Skip to main content

Full text of "Administrative control and executive action"

See other formats


Control and 


Control and 


Edited by 

B. C. ItMKE 


Professors of Business 

Michigan State University 


© 1961, by Charles E. Merrill Books, Inc., Columbus, Ohio. All 
rights reserved. No part of this book may be produced in any 
form, by mimeograph or any other means, without permission in 
writing from the publisher. 

Library of Congress Catalog Card No.: 61-8075 


The Book. The purpose of this book of over seventy readings is to pro- 
vide a balanced, concise, and varied source for the study of the many-sided 
aspects of control. It should appeal to those who must exercise control 
because of their responsibilities, and to those who must study and teach 
control because of its fundamental importance in administration. 

The subject matter of the large number of executive and management 
development programs and continuing education courses offered to em- 
ployees in business, government, and the professions, not only should but 
must include the study and discussion of elements of control because con- 
trol is an essential ingredient of successful management at all levek In 
fact, the proper exercise of control often is the key to managerial effec- 
tiveness and growth. . . 
College and university preparation for administration-whether in busi- 
ness or in the professions-must effect a thorough understanding of the 
function of control, of its varied applications and principles, and of the 
skills and techniques needed to control the activities of an enterprise. 

This book of readings is designed to serve the above areas by grouping 
carefully selected readings in the theory and application of control. Section 
A of the book introduces various definitions of control and continues by 
discussing the background of control, especially its important position in 
the emerging science of administration. Sections B, C, and D offer a variety 
of applications, including the newer tools of applied mathematics Finally 
; Section E presents a challenging and stimulating view of the future of 
administration, especially as it relates to control matters. 

This book is adapted for use as a major textual device in a course in 
control. It also is an invaluable supplementary device for a course built 
around lectures, discussions, cases, or problems which deal with specific as 

well as general areas of administration and control Attention is called to 
chapters which deal with the place in control matters of accounting and 
reporting, cost analysis, budgeting, break-even analysis, and other selected 
management tools. Some chapters discuss various planning and decision- 
making processes relating to both short- and long-run considerations of 
profitability and capital investment. Other chapters deal with the appli- 
cation of control to over-all administration, finance, production, market- 
ing, personnel administration, and more specific areas within these classifi- 

A book of readings provides the opportunity to present a variety of 
authors and, therefore, a cross-section of varied approaches. In this book, 
an article may combine several topics which are presumably dealt with in 
other articles or in separate chapters. 

In quite a few instances, portions of articles, some footnotes, and an 
occasional illustration have been omitted from the articles presented in 
this book in an attempt to minimize duplications and remove material 
irrelevant to the specific intent of this book. In no case was any attempt 
made to vary the position taken by the authors or to change the primary 
intent of their articles. 

Control. The increasing complexity and growth of modern organizations 
and the vastly improved methods of communication and reporting have 
forced management to pay greater attention to the administrative process. 
Control has shared in this development. Control follows planning, but 
just how closely it follows and how much it includes depends on how 
broadly or minutely the administrative process is subdivided. It occurs at 
all levels and in all types of activity, whether creative or routine, and 
whether human or mechanical. 

The word control when used negatively may imply purposeful restraint 
or restriction. When used positively, control includes the helpful aspects of 
review and guidance with strong overtones of supervisory powers. From 
an administrative point of view the minimum definition includes the re- 
view of actual progress by comparison with the plan and observation of 
the variance or deviation. An expansion of the definition adds the "doing 
something" about the pre-emergence, correction, or disposition of the 
variance or deviation. 

Control must be adapted to the specific administrative process of which 
it is a part. Thus, it is necessary in this book to discuss administration in 
a general sense in order to fit control into the proper perspective. Ad- 
ministration is presented in its broadest dimension as it applies to any 
economic, political, or social organization. After control has been identi- 
fied as a segment of administration, it is more closely scrutinized from 
varying viewpoints through the use of general and specific applications. 
The specific applications are by and large from the business field. 


The authors have received valuable suggestions, evaluations, and criti- 
cisms from various sources, particularly from their colleagues and students 
of various courses at Michigan State University. 

The encouragement given to the project by Dean Alfred L. Seelye and 
Associate Dean Kullervo Louhi, Graduate School of Business Administra- 
tion, is gratefully acknowledged, as is the advice and technical assistance 
of Eli P. Cox, Director of the Bureau of Economic Research, and Pro- 
fessors Edward M. Barnet, Gardner M. Jones, Dalton E. McFarland, 
David G. Moore, and Rollin H. Simonds, all of Michigan State University. 

Especially helpful in the early stages of the project were the suggestions 
of the authors of a successful book in Managerial Marketing: Associate 
Dean Eugene J. Kelley of the Graduate School of Business Administration, 
New York University, and Professor William Lazer of Michigan State 
University. The following persons rendered essential and devoted service 
in compiling and preparing the manuscript: Darwin J. Casler, James 
W.MacMeekin, LaDonna Michaelsen, Carolyn C. Mueller, and Daniel L. 

Special acknowledgment and thanks is due to Thomas N. Horton and 
Niels C. Buessem, Editors, Charles E. Merrill Books, Inc., without whose 
invaluable assistance and editorial guidance the book could not have been 
published. The cooperation of the various primary domestic and foreign 
sources of the materials included in this book are also gratefully acknowl- 



Section A The Elements of Control, 3 

James L. Peirce 
Robert Tannenbaum 

Alex W. Rathe 
Bradford Cadmus 

i The Planning and Control Concept, 

2 The Manager Concept: A Rational 
Synthesis, 16 

3 Management Control, 26 

4 Operational Auditing, 32 

5 Statement of Responsibilities of the 
Internal Auditor, 38 

6 What is a Controller?, 39 

7 Internal Control, 40 



Arnold F. Emch 8 Control Means Action, 43 
Robert L. Katz 9 Skills of an Effective Administrator, $2 
Robert M. Hutchins 10 The Administrator Reconsidered, 64 


Herrymon Maurer 
Robert Tannenbaum 

Richard M. Cyert 

Herbert A. Simon 

Donald B. Trow 

R. K. Gaumnitz 
O. H. Brownlee 


1 The Age of the Managers, 70 

2 Managerial Decision-Making, 81 

3 Observation of a Business Decision, 100 

14 Mathematics for Decision-Makers, 108 


Kenneth E. Boulding 

Herbert A. Simon 

Robert V. Presthus 

Edward H. Litchfield 

15 Evidences for an Administrative Science: 
A Review of the Administrative Science 
Quarterly, 118 

16 Comments on the Theory of 
Organizations, 124 

17 Toward a Theory of Organizational 
Behavior, 133 

18 Notes on a General Theory of 
Administration, lyi 

Section B The Major Areas of Administrative 
Control, 165 


A. F. D. Campbell 19 Cash Forecasting, 170 
Grover E. Edwards 20 Structure and Services of the Cash 

Budget, 1 7 4 
Maurice Moonitz 21 Reporting on the Flow of Funds, 182 
Bernard Whitney 22 A Working Model of the Financial 
Marion S. Israel Dynamics of a Business, 196 


Herbert J. Richmond 23 Streamlined Inventory Control and Stabi- 
lized Production Planning, 203 

John F. Magee . 24 Guides to Inventory Policy I. Functions 
and Lot Sizes, 214 

John F. Magee 25 Guides to Inventory Policy II. Problems 
of Uncertainty, 231 

John F. Magee 26 Guides to Inventory Policy III. Anticipat- 
ing Future Needs, 254 


Edward P. Burnham 27 Controlling the Costs of Research, 272 
Adolph G. Lurie 28 Controlling Research Costs with a 
Budget, 2j8 
Cyril C. Herrmann 29 Operations Research for Management, 286 
John F. Magee 



Donald J. Coppotelli 30 Departmental Results— We Get Them 

from Our Computer, 306 

J. Curran Freeman 3 1 Planning the Future of a Successful 

Product, 313 

Allen W. Rucker 32 "Clocks" for Management Control, 321 

Frank J. Andress 3 3 The Learning Curve as a Production 
Tool, 342 


T. V. Houser 34 The True Role of the Marketing 
Executive, 362 
Charles W. Smith 35 Planned Marketing: Key to Increased 

Profits, 370 

E. J. Hanley, et al. 36 Executives Tell How to Use Cost 

Accounting in Marketing, 376 


Thomas J. Wright 37 An Important Management Help: A 

Planned Program for Personnel Quality 
Control, 382 
Chris Argyris 38 Human Problems with Budgets, 390 
T. O. Yntema 39 Establishing More Effective Management 

Controls, 411 
RensisLikert 40 Measuring Organizational Performance, 


Section C Methods and Procedures of 
Executive Action, 432 


Philip A. Scheuble, Jr. 41 Horw t0 Fi Z ure E 4 ui P ment 

Replacement, 43$ 

Ezra Solomon 42 The Arithmetic of Capital-Budgeting 
Decisions, 456 

43 Capital Budgeting-How to Use it 
Constructively, 463 

44 How to Find the Moment When 
Modernizing Pays Best, 470 


Alfred R. Oxenf eldt 4 5 Multi-Stage Approach to Pricing, 483 

Walter B McFarland 46 How Standard Costs are Being Used 
Walter 15. ivicr ana u 4 ^^ ^ Contr(?h Bud geting, Pnctng: 

A Survey, 491 
Clarence B. Nickerson 4 7 The Cost Element in Pricing, 500 

R M Cyert 48 Organizational Structure and Pricing 
J. G. March Behavior in an Oligopolistic Market, $12 


Joel Dean 49 Measuring the Productivity of Capital, $ 1 8 
Leland G. Spencer 50 The Profitgraph-Technique and 

Applications, $3$ 
Gordon Shillinglaw 51 Guides to Internal Profit Measurement, 55/ 
Joel Dean 52 Decentralization and Intracompany 
Pricing, 570 


James D. Willson 53 Practical Applications of Cost-Volume- 
Profit Analysis, 583 

James D. Willson 54 Dynamic Budgeting— Getting the Most 

from Your Program, 597 

Edward W. Binshadler 55 Financial Planning and Control— The Key 

to Efficient Management, 601 
J. Fred Weston 56 Forecasting Financial Requirements, 60S 

Section D Control and the Accountant, 624 


Clayton W. Anderson 57 Disclosure of Assumptions— Key to Better 

Break-Even Analysis, 628 

A. W. Patrick 58 Some Observations on the Break-Even 
Chart, 634 

S. Alden Pendleton 59 Variable Budgeting for Planning and 

Control, 644 
Marjorie M. Osborn 60 Graphic Presentation of Results under a 

Variable Budget Control System, 654 


Henry J. Engler 61 The Accountant's Business with Business 
Forecasting, 661 

William J. Vatter 62 The Control Function of the Accountant 

as an Indispensable Part of 
Management, 666 
A. Charnes 63 Management Models and Industrial Appli- 
W. W. Cooper cations of Linear Programing, 67$ 


W. E. Harrison 64 

Fred G. Tuttle 

Alphonse Riverin 




The Contribution of Marginal Costing to 
Present-Day Problems, 701 

Dynamic Variable Cost Control, 710 

The Values and Uses of Distribution Cost 
Analysis, 720 

Tentative Statement of Cost Concepts 
Underlying Reports for Management 
Purposes, 727 

Direct Costing, 742 
Direct Costing, 747 

Section E Administrative Control and the 
Future, 756 


Wroe Alderson 70 
James A. McFadden , Jr. 71 

Peter F. Drucker 

Harold J. Leavitt 

Thomas L. Whisler 


New Concepts of Information for Man- 
agement Decisions— In Marketing, 758 
New Concepts of Information for 
Management Decisions— Research and 
Development, 764 

The Next Decade in Management, 770 
Management in the 1980's, 777 



Section A: 


Debate as to whether administration or management is an 
"art" or a "science" will undoubtedly go on endlessly; in 
part, but not altogether, this is due to a lack of common 
agreement on the differences between the two. Journals, 
such as Management Science and the Administrative Science 
Quarterly, by their very titles imply that a sizeable group 
regards "science" as a proper designation. 

Kenneth E. Boulding prepared a review of the first two 
volumes of the Administrative Science Quarterly, and that 
portion of it is included in the readings in which he gives 
his views on administration as a science. Theories are an es- 
sential ingredient of growth in any science, and articles by 
Edward H. Litchfield, Herbert A. Simon, and Robert V. 
Presthus propose theories of administration and organiza- 
tional behavior. Administration or management is dealt with 
in these articles as applicable in a broad but realistic sense to 
any organization, without any special regard for the nature 
or purpose of the organization. The implication of this 
should not be overlooked, for in one sense at least it means 
that a person skilled in administration can readily shift from 
one organization to another without automatically incurring 
the penalty of failure. It should not be assumed, therefore, 
that the articles just mentioned are limited to the business 

enterprise: this can also be said of a fair number of other 
articles included in this book of readings. 

Perhaps those who prefer to regard administration as an 
"art" may find some indirect support in the article by 
Robert M. Hutchins in which he candidly describes por- 
tions of thirty-three years of administering non-profit or- 
ganizations including some overtones of how he would do 
it now if he had to do it over again, based on experience 
and hindsight. 

Regardless of the classification favored by the reader, 
there is little doubt but that administration or management, 
of business especially, has tended to evolve as a clearly 
defined profession in the United States because of the exten- 
sive use of the corporate form of organization which encour- 
ages a marked separation of ownership and management. 
The growth in size and respect of collegiate schools of 
business reinforces the concept of business administration 
as a distinct professional area. 

The remainder of the articles in this section tacitly as- 
sume, for the most part at least, that it is business which is 
being discussed either directly or indirectly. The growing 
importance of business management is stressed by Herry- 
mon Maurer in "The Age of the Manager." The importance 
of management in general and its important function of 
decision-making are thoroughly discussed and outlined in 
two separate articles by Robert Tannenbaum. Robert L. 
Katz also gives an indication of the scope of business man- 
agement by reviewing the several skills which an effective 
administrator should possess. Richard M. Cyert, Herbert A. 
Simon, and Donald B. Trow are co-authors of "An Ob- 
servation of a Business Decision." R. K. Gaumnitz and 
O. H. Brownlee are co-authors of "Mathematics for De- 
cision Makers" which opens a new array of management 
aids made possible by the application of mathematics which, 
in turn, is made feasible by new hardware such as the elec- 
tronic computer. Other articles relating to mathematical 
applications to specific problems appear in other sections of 
this book of readings. 

Control, another term which is subject to a variety of 
definitions, is one of the extremely important functions of 
management which follows closely after decision-making. 
Control may be viewed as playing the relatively passive 

role of reporting the progress made in carrying out a de- 
cision once it is made, usually by comparing the plan which 
implemented the decision with the actual accomplishment 
in order that deviations or exceptions can be noted and ap- 
praised. Control can also be assumed to play an active role 
by adding to the foregoing that control should also include 
the purposeful activities of assuring compliance with the 
plan. Arnold F. Emch emphatically takes the latter position 
in "Control Means Action." Articles by Alex W. Rathe and 
James L. Peirce continue in much the same vein by dis- 
cussing control as a tool for carrying out plans incorporat- 
ing, as they do, organizational principles and plans which 
they regard as important for such purpose. 

An article by Bradford Cadmus on "Operational Audit- 
ing" discusses a relatively new use or extension of internal 
auditing into management areas. Internal auditing, in turn, 
is described in the "Statement of Responsibilities of the 
Internal Auditor" as issued by The Institute of Internal 
Auditors. Internal control is defined in an excerpt from a 
pamphlet issued some years ago by the American Institute 
of Certified Public Accountants. The staff position of con- 
troller is ideally and realistically described by an excerpt 
from a recent publication of the Controllers Institute of 
America. When finished with this section, it will be ap- 
parent to the reader that the word control has a central 
meaning, to be sure, but also a number of interesting facets. 



James L. Peirce* 

Mr. James L. Peirce uses his definition of planning and control to 
differentiate between the proper authorities and responsibilities of the 
president and other operating heads of a business corporation and those 
of the controller. The budgetary process is regarded as an essential ele- 
ment in both planning and control. 

A discussion of the planning and control idea might be started by say- 
ing that it is synonymous with management itself. Certainly no business 
can exist without some form of this twin concept, and it might perhaps 
be demonstrated that success in business is proportionate to the astuteness 
of its planning and the skill with which it is controlled. 


It may be useful at the outset to define "planning" (as the word is 
understood in the concept we are discussing) and to place it in relation 
to the control function. Planning, of course, is carried on during every 
hour of the business day, and sometimes during many other hours besides. 

* From The Controller, XXII, 9 (1954), 403-6, 422, 424-25. Reprinted by permis- 
sion of The Controller. 

g What Is Control? 

It may exist with or without control-that is, with or without disciplined 
efforts to follow the plan or to explain deviations from it. On the other 
hand, control cannot exist without planning, and therefore the planning 
must be designed to fit the specifications of control. 

In the modern sense of an integrated planning and control system, then, 
planning refers to the construction of an operating program, comprehen- 
sive enough to cover all phases of operations, and detailed enough that 
specific attention may be given to its fulfillment in controllable segments. 
It may therefore be reiterated that the planning process must be con- 
ducted in direct relation to the needs of control. 


An examination of this word "control" shows that, like many another 
word in the English language, it has a number of meanings. It is absolutely 
necessary to have a clear understanding of its definition in the specialized 
sense in which it will be used herein. 

Perhaps the easiest approach to this shade of meaning is to mention 
some of the things that the word "control" in this usage does not include. 
It does not signify the kind of control over a business enjoyed by a ma- 
jority shareholder. It does not refer to any part of the control centered 
in a board of directors or a president. It does not include line authority 
for making or carrying out policy or operating decisions. 

What then is "control" in the sense to be used here? It is defined as 
the presence in a business of that force which guides it to a predetermined 
objective by means of predetermined policies and jteisions. Every busi- 
ness executive can identify this control force in his company. It operates 
quite apart from the mass of operating decisions and instructions con- 
stantly emanating from the line organization. It does not steer the course, 
but it informs operating management at_once of any significant deviation 
from it. It does not take action, but it frequently impels action by turning 
a spotlight on the pertinent facts. 


The practice of this variety of control may be referred to as controller- 
ship. When delegated, it is exercised by an executive properly called a 
controller, although he may not actually carry that title. His title may 
be, for example, vice president and treasurer and his controllership assign- 
ment intermingled with financial and other administrative duties. Never- 
theless, regardless of title, because he performs the controllership func- 
tion, he will be referred to here as the controller. 

The Planning and Control Concept p 

Planning is the primary duty of the president, assisted by all line and 
staff executives. The control function is exercised by the same group, 
but may be centered functionally in the controller. The objective of con- 
trollership is to assist all levels of management in controlling to the plan. 
It never issues orders, but it coordinates the machinery of planning, re- 
cords and reveals the facts and makes plain to those in charge what they 
must do to achieve the prescribed aim. 


The control process, like all effective modes of management, rests on a 
simple principle. It may be stated as a three-part formula: 

The first component is the adoption of a plan. 

The second is reporting actual performance as compared with the plan. 

The third is making decisions and taking action. 

This pattern is repeated as many times as there are units of responsible 
supervision in the company, and the whole is assembled as a grand plan 
which directs the company's operations and controls its course. As time 
goes on, experience sometimes dictates alterations in the plan. Action 
taken as a result of reporting performance against plan reaches forward 
to new and better plans. 

It should be evident that the three phases of the formula are operating 
concurrently— that a management following this system will be constantly 
planning, reporting and taking action. The important thing to note is 
that the decisions reached and the actions taken will be directly related 
to a master plan. They will not be spasmodic nor will they be consum- 
mated without adequate reference to the fundamental objectives of the 
business. No decision or action will be taken which is out of harmony 
with actions in other departments of the business, because all are gov- 
erned by a universal plan. 


The initial step, the adoption of a plan, is perhaps the most difficult. 
Even after the habit of planning has become ingrained, it is not easy to 
induce a group of executives to set aside pressing matters and think into 
the problematical future. 

First, consider the form in which the planning is to be stated. The 
common denominator, of course, is money. All planning must ultimately 
be translated into dollar figures, which is the language in which business 
operates. The ultimate form of the programming therefore is, typically, a 
planned profit-and-loss statement for a forthcoming period of, say, 12 

IO What Is Control? 

months, supported in detail by sales budgets or forecasts, expense budgets 
and so on, and also supplemented with detailed explanations. The assump- 
tions, bases and computations upon which the budget figures are predi- 
cated all should be recorded because they will later serve a vital purpose. 

These figures are not the plan itself. They are only the external ex- 
pression of the plan, in a language understandable to all. The statements 
are a mere vehicle with which to inform, appraise and perchance read- 
just. The substance of the plan itself is in the minds of its creators. 

It would hardly be necessary to emphasize this self-evident truth if it 
were not ignored so often. It is of no value whatever to budget a given 
amount for advertising, for instance, without clear advance knowledge 
on the part of the advertising manager of the media to be used, the 
markets to be reached, the products to be advertised. In fact, the adver- 
tising budget does not qualify as a segment of a true operating plan un- 
less the sales manager also understands the exact degree of support it will 
afford him in selling his forecast volume. 

To be effective, planning cannot be superficial. It depends on a firm 
statement of principles by the top executive, a clear understanding by 
each man of the contribution his division or department is expected to 
make to the enterprise, and a willingness to plan with care and to stand 
back of the plan. 


Anyone who has taken part in the preparation of a plan of operations 
for an industrial company is aware that the undertaking is far more com- 
prehensive than it may first appear. No activity of the company is ex- 
empt; every segment must be fitted into a master program. 

Consider for a moment the implications of preparing a sales budget, 
sometimes referred to as a sales forecast, for a period of a year ahead. All 
products must be budgeted, including those which have not yet been 
introduced. Due weight must be given the general economic outlook and 
its bearing on the demand for the company's products. The effect on 
volume of proposed changes in merchandising methods must be con- 
sidered. Both volume and selling prices must be planned, and this involves 
an advance determination of the quantities to be sold in each market or 
through each sales outlet. 

Bear in mind that creative sales budgeting will not tolerate retrospec- 
tion, astrology nor guesswork. The penalties are too severe. For example, 
the sales promotion budget will lean heavily on the planned sales volume. 
What is perhaps even more significant, so will factory production levels, 
purchase commitments and the ever-critical acquisition and layoff of 
production workers. Planning is a company-wide process of integration, 
in which no man can stand alone. All depend upon each other. 

The Planning and Control Concept j x 

When the volume of planned sales has been established, the manufac- 
turing division is in a position to plan production levels and times, as well 
as inventories. Manufacturing costs must then be fitted into the program, 
including material purchase prices, wage levels and even manufacturing 
efficiency. The planning process then fans out to include factory adminis- 
tration costs, research objectives, selling strategy and general administra- 
tion. The spotlight is turned on the future course of each activity and 
each is reduced to budgeted figures. 

Finally, when all this planning has been done and translated into the 
language of dollars— after financing decisions have been made and esti- 
mates of money costs and income taxes prepared— it is possible to arrive 
at the planned net profit. It is this figure which determines return on the 
capital invested in the business and willingness on the part of investors to 
provide more capital as required by its growth. 


We have skimmed lightly over a process which sometimes taxes the 
capacity and judgment of every executive in the business. It is not an 
easy task to make major operating decisions for 12 months ahead and 
place them neatly on a timetable. Yet this is what must be done if the 
company is to operate under the kind of control that produces satisfac- 
tory results. 

Sometimes these advance determinations are subject to subsequent re- 
vision. The planning process should accommodate this need. If decisions 
must be made subject to probable change, all concerned should under- 
stand the assumptions on which they are based and the extent to which 
the plan may have to be altered. Flexibility, not rigidity, is a characteris- 
tic of dynamic planning. 

The dividend from this process is proportionate to the magnitude of 
the hurdles surmounted. First, there is time for deliberation on the prob- 
lems appearing on the horizon, as contrasted with solving them at the 
last minute in an atmosphere of crisis. And second, the kind of inter- 
change described in the foregoing illustration yields a mutual under- 
standing of the basics of the business which could not be obtained in any 
other way. There is no comparable training course for executives. 


Planning of the company's operations cannot be considered complete 
unless it is integrated wtih a plan for providing required new buildings, 
machinery, equipment, tools and so on. The needs of the organization in 
this area should be assembled as carefully as are the departmental operat- 
ing budgets. 

What Is Control? 

In this case, however, the term of planning is usually a little longer- 
say three years, as compared with the customary 12-month projection of 
sales and expenses. Capital requirements must be prepared on a long- 
range basis because of the time requirement for construction and procure- 
ment and because these items are largely charged against the income of 
future years. 

Furthermore, such expenditures will largely govern the planning for 
financing. The logical outcome of this thinking is a phase of planning 
which is far broader than capital expenditures alone, and which might be 
referred to as a financial program. 

In constructing such a program, estimates must be made of required 
operating cash, accounts receivable and inventories, as well as plant and 
equipment items for the planned period. In fact, every item on the com- 
pany's balance sheet receives scrutiny in this process, and the resulting 
financial program may be expressed in terms of pro forma balance sheets 
for each of the succeeding three years. Obviously, borrowings, net 
profits, dividends and enlistments of new equity capital must be planned 
in order to accomplish this result. 

The process of constructing a financial program is perhaps even more 
difficult than that of creating an operating plan for a one-year period. It 
requires that proposed major moves be frozen into at least tentative de- 
cisions, which can usually be made only by boards of directors. 


It will be recalled that the control formula consists of (1) the adoption 
of a plan, (2) reporting actual performance as compared with the plan, 
and (3) making decisions and taking action. The second of these steps 
merits a brief discussion. 

Probably the simplest known form of reporting performance against 
plan is the typical expense report issued by the accounting department, 
showing itemized expenses for the month, compared with budget figures. 
Too often the reporting ends there. It should add two other features: a 
written explanation of the figures where it would be helpful, and sym- 
pathetic consultation with the recipient of the figures. The latter should 
only be lengthy enough to ascertain that the man responsible really under- 
stands the meaning of the figures. 

The same principle, of course, applies to profit-and-loss statements and 
to reporting on the performance of any unit of the business. The report- 
ing cannot be perfunctory. It must be based on an intimate knowledge 
of the operation and of the plan itself, rather than merely on the figures. 
For instance, it is far more significant to point out a deviation from a 

The Planning and Control Concept 


planned expansion of the sales force than to report only that salaries are 
so many dollars less than the amount budgeted. 


The final step in the control formula is decision and action. When the 
planning has been done properly and adequate reporting has been made 
on performance against the plan, the ensuing decisions sometimes become 
surprisingly clear. The action is frequently indicated in the reporting 

Assume, for example, a failure of actual manufacturing cost to match 
the planned cost for a given month. The result, of course, shows up in a 
deficient net profit. The excess of actual over budgeted or planned cost 
has been traced to its source. It is relatively easy to do this if the plan 
has been constructed in adequate detail by the manufacturing organiza- 
tion. Whether the reason be low production volume, high material prices, 
heavy waste losses, or any of the myriad of other happenings which push 
costs upward, management is faced at once with a clear-cut decision. 
Either the condition must be corrected, or, if this is not possible, other 
changes in the plan must be made to compensate for it. Planned expenses, 
costs or sales volume must be improved, or planned net profit must be 

The decision should be reached in an atmosphere of participation by 
everyone concerned. If an adjustment is made in the plan itself, each re- 
sponsible executive and supervisor accepts the full impact of the change 
on the performance expected of him. 

In practice, an orderly method is required for revision of the operating 
plan and reflection of the change in the projected operating figures. For 
example, it may be advisable to give effect to new planning and changes 
in plan only at three-month intervals, in the course of a complete revision 
of budgets. Interim deviations, meanwhile, are made conspicuous, in 
harmony with the best management-by-exception tradition. 

By these devices of control, all units of the business are coordinated. 
The simple triad, planning-reporting-action, becomes the guiding prin- 
ciple of the business. 


It is a long step from the resolution to have better planning and con- 
trol to its actual realization. As in every other advance in management 
method, responsibility must be fixed at the outset. In this case, however, 
the assignment of responsibility is decidedly complex. 

Iz j. What Is Control? 

Fundamentally, every manager or supervisor in charge of a unit of the 
business well enough defined to have a budget of its own must be made 
responsible for the planning and control of that unit. By implication, the 
responsibility travels up the organization line all the way to the top. 

The president, of. course, is ultimately charged with the obligation of 
success in this field, as in all others, and he therefore must undertake to 
see that the control mechanism is constructed and maintained and that 
the entire organization is educated in its use. 

Because this multiple task is so time-consuming, experience has proven 
the wisdom of assigning it to a staff executive who may be referred to 
as the controller. 

Essentially the burden of installation, education and follow-up falls on 
the president and the controller. The respective areas of action of these 
two executives require a little further comment. 


It is axiomatic that all policies of management must enjoy the unquali- 
fied support of the top man in the business. Planning and control tech- 
niques are no exception. The company's president must understand 
them, use them himself and furnish the required leadership in their ap- 

It is probably self-evident that the same comments apply, in degree, to 
executive vice presidents, division and department managers and others. 
The acceptance and use of the planning and control concept must be 
commensurate with the authority invested. It is the president's task to 
create an understanding of these points in his immediate subordinates, 
who in turn are held accountable for transmitting this understanding and 
making control effective throughout their respective spheres of activity. 


It should be re-emphasized that the control function, within the con- 
cept we are discussing is not always assigned to an executive with the 
title of controller. It is frequently found to reside in a top financial or 
administrative officer, and sometimes remains in the hands of the chief 


As defined by the Controllers Institute of America, the functions of 
controllership include establishing, coordinating and maintaining an in- 
tegrated plan for the control of operations, but it is specified that this 
must be done through authorized management. Such a plan, it is stated, 
would provide cost standards, expense budgets, sales forecasts, profit 

The Planning and Control Concept 1 5 

planning and programs for capital investment and financing, together 
with the necessary procedures to effectuate the plan. The important 
words in this assignment are "through authorized management" and this 
little phrase sets the keynote for the controller's peculiar mode of getting 
things accomplished. He himself should never establish a single standard 
or budget (except his own), nor a single sales forecast. The plan must 
be constructed, under his helpful guidance, by the operating executives 
who will have to accept the responsibility for performance. 

The Institute's definition then assigns to the controller the duty of 
measuring performance against approved operating plans and standards, 
and of reporting and interpreting the results of operations to all levels 
of management. It is within this assignment that he finds the need for 
designing, installing and maintaining accounting and cost systems and 
records, determining accounting policy and compiling statistics. 

Other parts of the definition equip the controller with power to 
measure, interpret and report on almost anything— even the validity of 
the objectives of the business— and to consult with all responsible seg- 
ments of management on any phase of the operation of the business. He 
is also charged with the duty of interpreting economic and social influ- 
ences in their impact on the business. He is free, in fact, to offer his 
constructive thinking wherever he feels that it will contribute to more 
effective planning, direction, control. 

A discreet controller can exercise this wide latitude without offense to 
his fellow executives, provided he heeds the ground rules of controller- 
ship. These include a fastidious abstention from taking on operating 
responsibilities or making operating decisions; reporting consistently to 
all concerned; insisting that the line organization determine their own 
budgets and standards of performance; and above all, reporting and 
interpreting without exaggeration, bias or regard for the preconceptions 
of others. 

In particular also, he must not take operating people to task for failure 
to meet the standards, and he must not be placed in the position of 
making negative decisions on spending money. It is a familiar myth that 
these unpleasant attributes are the characteristics of a controller. The 
penalty for permitting this misconception to be accepted among his 
associates is the sacrifice of controllership effectiveness. 

The particular state of mind that the controller ought to impart to the 
organization is a sense of balance, stability and direction. 


It would be unfortunate to leave the impression that planning and con- 
trol solve all the problems of running a business. Planning and control 

1 6 What Is Control? 

simply facilitate the solution of these problems very materially and open 
up possibilities of achievement which could not be realized otherwise. In 
a word, it represents thinking forward instead of meeting each daily crisis 
when it arises. It represents detailed knowledge of where and why we are 
going astray, as contrasted with a tardy awakening to developments 
which have been buried for too long in the debris of current affairs. Its 
fruition is a priceless sense of knowing where we are going rather than 
steering a blind course. 

This is a present possibility. We are just beginning to learn the un- 
limited benefits of wisely conducted planning and of control construc- 
tively applied. New and better methods will constantly improve the prac- 
tice of planning and control, but its unchanging principles, symbolized 
by the planning-reporting-action formula, are the eternal possession of 
business management from the time of their discovery. 


Robert Tannenbaum* 

Professor Robert Tannenbaum defines a manager as an individual who 
"has and uses formal authority to organize, direct, or control responsible 
subordinates?'' In the following excerpt from a larger article he discusses 
the scope of this definition. A review of the literature pertaining to man- 
agement, which forms a large section of the total article, has been omitted 
in the following material. 

An enterprise may be viewed as an instrument for the transformation 
of the services of persons and things into completed product. Of the 
personal services contributed to an enterprise, some are managerial in 
character; others, nonmanagerial. Those who contribute managerial serv- 
ices will be called managers, while those who contribute nonmanagerial 
services (although usually called workers or laborers) will be called non- 

This article is primarily concerned with the nature of managerial serv- 
ices. The problem is in clearly differentiating them from nonmanagerial 
services. Such differentiation can best be accomplished by isolating those 
functions performed exclusively by managers. 

*From The Journal of Business, XXII, 4 (1949), 225-41. Reprinted by permission 
of the University of Chicago Press. 

The Manager Concept 1 7 

Taking into account the definitions employed and differing termino- 
logical usages, eliminating ;n most cases terms which represent processes, 
techniques, or tools, and allowing for ambiguities, it is possible to classify 
into five groups the various functions discussed by "the writers. Each 
group is characterized by the fact that (a) the functions of which it is 
comprised are the same or similar but are assigned different names or (b) 
that the functions are different but are closely related or (c) that they 
overlap in content 

In the first group are "organization," "lay out the broad lines of ad- 
ministrative structure," and "develop and maintain a system of com- 
munication which jointly involves a scheme of organization and an 
executive personnel." In the second group are "initiation and approval of 
decisions," "planning," "formulate and define the purposes, objectives, 
ends, of the organization," "formulation and determination of policy," 
and "direction." In the third group are the terms "control," "supervision," 
and "appraisal." In the fourth group are "inspiration," "motivation," 
"leadership," and "promote the securing of personal services." And in the 
fifth group are "trusteeship" and "representation." Finally, it should be 
noted that "co-ordination" appears as a function in many of the quota- 

This grouping of the functions of managers as presented by various 
writers provides a convenient basis for the development of a synthesis of 
the manager concept which is attempted in the following section. 


The functions in general- -The functions of managers may now be 
listed and discussed in detail. No claim to complete originality is made 
for this presentation. It represents, for the most part, an effort logically 
to combine selected ideas of many writers into a meaningful and useful 
functional definition of the manager. 

No special brief is held for the terminology chosen or for the particu- 
lar grouping of activities used. 

In speaking of the functions of managers, it is not intended to imply 
that each manager in an enterprise performs all the functions. Such is 
seldom, if ever, the case. Managers, like nonmanagers, are specialists. They 
typically specialize in specific functions or a specific function. The func- 
tions of managers are those performed exclusively by managers as a group. 

It is the thesis of the present discussion that all managerial activities 
are included in three functions: organization, direction, and control. 
These are derived from the groupings presented at the conclusion of the 
preceding section and will now be discussed. 

1 8 What Is Control? 

Organization- The term "organization" implies an arrangement in 
which all units are so related to each other that they may work as a 
whole, each unit having its proper task to perform, 1 and "to organize" 
means "to arrange or constitute in interdependent parts, each having a 
special function, act, office, or relation with respect to the whole." 2 
These statements include two basic concepts, namely, units or parts each 
having its proper or special task to perform and an arrangement involv- 
ing an interdependence or relationship between the units or parts. The 
managerial function of organization involves these two concepts. Man- 
agers must determine the degree and type of specialization to be effect- 
uated within the enterprise and they must determine the relationships 
that are to exist among the specialized units. 

With respect to the degree and type of specialization, one of the char- 
acteristics of the individuals, groups, and complexes comprising an enter- 
prise is that they each contribute specialized services to the group or 
complex of which they are a part. The determination of these specializa- 
tions involves analysis first and then synthesis. 

The function of organization begins with the objective of the enter- 
prise, i.e., with the good or service to be produced. It must be determined 
by analysis what services of individuals will be necessary to produce the 
good or service in question. This determination entails questions relating 
to both degree and type. How specialized should be the services to be 
contributed by each individual occupying a position, and what should be 
the type of these services? Both of these questions must be answered. 
When they are, the process of synthesis can begin. First, the individuals 
contributing specialized services must be combined into groups. The 
nature of these groups is similarly determined by the degree and type of 
group specialization desired. Next, groups are combined into complexes 
and these into superior complexes and so on until the supreme complex 
is achieved: and always the degree and type of specialization are the de- 
terminants of the nature of each of these units. 

The services contributed by the managers who comprise the managerial 
superstructure are also specialized; and the determination of the degree 
and type of the specialization of these services must be made by man- 
agers. Certain aspects of this determination lead to a consideration of the 
second concept involved in the managerial function of organization, 
namely, the determination of the relationships that are to exist among the 
specialized units in an enterprise. 

1 Cf. "Order," Webster's Dictionary of Synonyms (Springfield, Mass.: G. & C. 
Merriam Co., 1942). 

2 "Organize," Webster's New International Dictionary (2d ed.; Springfield, Mass.: 
G. & C. Merriam Co., 1924). 

The Manager Concept 1 9 

The relationships established among the managers of an enterprise deter- 
mine the relationships among the groups and complexes which they 
head. So it is upon the former relationships that attention must be 
focused. The managerial relationships are always expressed in terms of 
authority and responsibility, and they are established by delegation. 
Therefore, each of these three concepts must first be defined. 

Authority is the right to command or to act. Thus, a person having 
authority has the right not only to act himself but also to expect action 
of others. But what is the source of this right? In practice, authority 
appears to originate at the top of a structural hierarchy— under private 
enterprise, with the owners— and to flow from owners to their represen- 
tatives, the managers, and from superior managers to their subordinates. 
Hereafter in this presentation, authority, when viewed in this customary 
manner, will be referred to as formal authority. 

Responsibility involves being subject to another who may exact redress 
in case of default. Responsibility is answerability or accountability. One 
is typically responsible to another for the performance of tasks assigned 
to him by the latter. 

Delegation is the act of investing with formal authority to act for 
another. "Delegation always means the conferring of authority, and can 
never mean anything else." 3 A delegation of formal authority must always 
include a definition of the limits within which that authority may be 

As has been indicated, the fountainhead of all formal authority in a 
private enterprise is the owners— in a corporation, the stockholders. The 
latter typically retain some formal authority but delegate most of it to 
their elected representatives, the board of directors. The board, in turn, 
becomes responsible to the stockholders for exercising the delegated 
formal authority within the specified limits. The board retains some for- 
mal authority and delegates the balance to the manager who heads the 
supreme complex. The delegation establishes the specialization for this 
manager, and he becomes responsible to the board within the limits of 
the delegation. This process continues downward through the managers 
of superior and subordinate complexes to the managers of groups. The 
latter delegate to the individuals comprising their groups formal authority 
to perform designated tasks, and the individuals become responsible for 
such performance. These individuals are never delegated formal authority 
to command, nor are they able to delegate authority to others. 

A manager who delegates formal authority to subordinates does not 
thereby escape responsibility to his superior for the exercise of the 
formal authority which the latter delegated to him. He is able to assume 

3 James D. Mooney and Alan C. Reiley, The Principles of Organization (New 
York: Harper & Brothers, 1939), p. 17. 

20 What Is Control? 

that responsibility by holding his own subordinates responsible for the 
formal authority which he has delegated to them. 

The formal authority which is delegated to subordinates may itself be 
specialized into the authority to prescribe and the authority to enforce. 
The former is authority to indicate how designated activities shall be 
performed, the latter is authority to see that the activities are performed; 
some managers exercise both types of authority, others, only one type. 
Formal authority may also be centralized or decentralized. The more 
centralized the authority, the more it has been reserved for execution 
by managers at the higher levels in the managerial superstructure; the 
more decentralized the authority, the more it has been delegated to the 
managers at the lower levels in the managerial superstructure. 

The process of delegation establishes definite relationships between 
managers and therefore between the specialized groups and complexes 
which they head. These relationships are those of superior and subordi- 
nate. A subordinate is always responsible to a superior for the accom- 
plishment of that for which he has been delegated formal authority. 
Formal authority is delegated downward through the managerial hier- 
archy; responsibility extends upward through the same hierarchy. The 
superior-subordinate interconnections are the channels of formal com- 
munication, both downward and upward, within an enterprise; and the 
managers are themselves the centers of communication. 

Direction— Once managers have determined the degree and type of 
specialization to be effectuated within the enterprise and the relationships 
that are to exist among the specialized units, they have provided them- 
selves with a mechanism for the attainment of purpose. They must next 
employ the mechanism. The first function of managers involving such 
employment is the function of direction. Direction is the use of formal 
authority in order to guide subordinates. Direction involves the devising 
of the purposes of action and the methods or procedures to be followed 
in achieving them. The decisions to be made in connection with direction 
must answer the questions "what?" "how?" "when?" and "where?" 

The devising of the purposes of action provides the "what-content" 
of direction. It has already been seen that the individuals comprising 
groups, the groups comprising complexes, and the subordinate complexes 
comprising superior complexes must in each case have an enterprise 
purpose, end, or objective. In addition, each individual manager and non- 
manager, has a purpose to achieve in his own activity. Managers must 
formulate these purposes for their subordinates and order them put into 

The Manager Concept 2 1 

The devising of purposes begins with the broad purpose or purposes 
of the enterprise. These are then translated into subpurposes for the 
superior complexes comprising the supreme complex. The subpurposes 
are further subdivided for the subordinate complexes - and so on down 
the structural hierarchy until each individual has his own purpose. These 
translations or subdivisions are made successively by managers, starting 
with those at the top of the hierarchy of the managerial superstructure 
(typically the board of directors) and moving down to those who head 

The devising of methods or procedures to be followed in achieving pur- 
poses provides the "how-," "when-," and "where-content" of direction. 
Here, again, the broad and general decisions are made by managers at 
the top of the managerial hierarchy, and these decisions are made ever 
more specific by successive subordinates down through that hierarchy. 

Directive decisions, once made, serve as a basis for the guidance of 
action. The vast majority of directive decisions are made to guide sub- 
ordinates in actions which are repeated frequently. Relatively few such 
decisions are made to guide actions which are performed but once. In 
the case of any action frequently repeated, a tremendous burden would 
be placed on managers if a duplicate decision had to be made each time 
the action were to be repeated. To avoid this unnecessary duplication 
in decision-making, managers have developed numerous devices or tools 
to be used in providing guidance for repetitive action. In practice these 
devices are variously referred to as "budgets," "policies," "procedures," 
"practices," "methods," "rules," "regulations," "routines," "schedules," 
"instructions," "specifications," "designs" etc. 4 The importance of these 
devices to managers cannot be overly stressed. Because they obviate 
the necessity for redeciding questions, they release for other purposes 
much valuable time which otherwise would have to be devoted to such 
redecision. These devices are also used by managers as criteria of action, 
since each of them implies a standard of performance to be attained. 
Serving as a guide to action and a criterion of action are simply two 
aspects of the same thing. 

4 Some of these terms are by no means mutually exclusive from the point of view 
of definition. Furthermore, in practice they are often used to refer to different 
things. There is a crying need for standardized terminology here. The term "policy" 
provides an excellent example of this need. It has been used in so many ways that 
it is necessary for each user to define the term in order for it to have any precise 
meaning in the context in which it is used by him. See Chester I. Barnard, "Com- 
ments on the Job of an Executive," Harvard Business Review, XVIII, No. 3 (spring, 
1940), 296. 

22 What Is Control? 

Control— The second function of managers involving the employment 
of the mechanism for the attainment of purpose (the organization) is 
the function of control. Control is the use of formal authority to assure, 
to the extent possible, the attainment of the purposes of action by the 
methods or procedures which have been devised. The execution of this 
function involves the selection and training of individuals, the provision 
of incentives, and the exercise of supervision. These components of the 
function of control (selection and training, incentives, and supervision) 
may appear at first glance to be unrelated activities. They do, to some 
extent, involve different managerial techniques. However, each is essential 
to the attainment of purpose by the methods or procedures which have 
been devised and is therefore logically classified under the function of 
control as defined. 

One aspect of the function of organization previously discussed is the 
determination by managers of the degree and type of specialization to 
be effectuated within an enterprise. In part, this determination results 
in specifications of the types of services which will be required of indi- 
viduals. Managers must next match these specifications with individuals- 
managers and nonmanagers— able to contribute the desired types of serv- 
ices. Such individuals may be found either within or without the 
enterprise. If they are found, they may be selected to fill positions calling 
for the types of services they are able to contribute. If such individuals 
are not found, then other individuals, either from within or without the 
enterprise, with the capacity for contributing the desired services must 
be selected and then trained until their capacity becomes ability. 

The task of matching individuals with specifications is not an easy one. 
An individual's ability to contribute the desired types of services is often 
closely related to such intangible personal factors as his character, per- 
sonality, temperament, and the like; and when individuals work together 
in co-operative groups, these factors are important determinants of inter- 
personal compatibility. Since no completely adequate measures of these 
factors have as yet been devised, the selection of individuals requires the 
exercise of judgment on the part of the manager making the selection. 

Now, each manager is responsible to his superior for the accomplish- 
ment of assigned tasks; and since the ability of a manager to meet such 
responsibility depends in part on the quality of his subordinates, and since 
the determination of that quality is based to a greater or less extent on the 
exercise of judgment, he must be able to select his own subordinates. In 
no other way can he reasonably be held for their performance. The selec- 
tion of subordinates is particularly crucial from the point of view of con- 
trol when those subordinates are themselves managers. This is true because 

The Manager Concept 2 3 

the intangible personal factors play such an important role in their work. 

By selecting subordinates and training them when necessary, managers 
try to provide themselves with individuals able to contribute the types of 
services necessary for the attainment of purpose. Any single manager can 
reasonably be held responsible for such attainment only if he has been the 
one whose judgment has determined the ability of his subordinates. 5 

It is not enough that individuals be found who are able (or who can be 
trained to be able) to contribute desired services to the enterprise. They 
must also be willing to do so. Ability must be supplemented by strong 
motivation. Unlike the flow of services from a machine, that from an in- 
dividual is subject to considerable variation in intensity through time de- 
pending upon the motivation of the individual; thus it becomes necessary 
not only to make individuals willing to contribute desired services but to 
regulate as far as possible the intensity of the flow of the services. Incen- 
tives must be provided for these purposes. 

An incentive as here viewed is any device which is offered to induce an 
individual— manager or nonmanager— to contribute services at a desired 
intensity to an enterprise. The inducements which may be offered to moti- 
vate an individual are numerous. They include various materials things: 
opportunities for distinction, prestige, personal power, and the like; de- 
sirable physical conditions of work; pride of workmanship, sense of 
adequacy, feelings of altruism, loyalty, etc.; social compatibility; custom- 
ary working conditions and conformity to habitual practices and attitudes; 
opportunity for the feeling of participation in the course of events; soli- 
darity or satisfaction of the gregarious instinct; and coercion. The proper 
use of incentives by a manager is a method by which he may secure and 
regulate the service contributions of subordinates that are so essential to 
the attainment of the purpose for which he is responsible. 

Individuals who are able and adequately motivated to contribute serv- 
ices may still, for many reasons, execute commands imperfectly. It will be 
recalled that directive decisions are often expressed in terms of criteria of 
action or standards of performance. The observation of performance, the 
comparison of it with the predetermined criteria or standards, and the tak- 
ing of remedial steps where called for are essential if the purposes of action 
are to be attained by the methods or procedures which have been devised. 

5 Many writers view selection as being a phase of the function of organization. To 
these writers, organization includes both structural and staffing considerations. A 
good case can be made for this point of view. I prefer, however, to include selection 
within the function of control. To me, the right to select subordinates— particularly 
managerial subordinates— is so essential to an assurance of the attainment of purpose 
that selection seems most properly classifiable as a phase of the managerial function 
of control. 

2 . What Is Control? 

These entail the exercise of supervision. 6 Supervision involves overseeing, 
inspection, the use of accounting and statistical devices, the use of reports, 
etc., for the purpose of determining the facts of performance; and it in- 
volves appraisal or evaluation for the purpose of comparing performance 
with standards. It is important to recognize that supervision is exercised 
not only by the managers of groups but by all managers who have 
subordinates, including the manager of the supreme complex and the 
board of directors. 

A comment on the managerial technique of command.- Command is a 
managerial technique used in connection with the execution of all the 
functions of managers. A command is an order from a superior to a 
subordinate to do something. Through command organizational, directive, 
and control decisions can be translated into action. Command, therefore, 
is (along with decision-making) probably one of the most important and 
pervasive of the managerial techniques. 

Additional "functions" considered. -In the classification of functions 
based on the formulations of other writers "trusteeship" and "representa- 
tion" comprised the fifth group of functions. In the writer's opinion, the 
so-called "function of representation" is not one of those functions per- 
formed exclusively by managers, nor can it serve as a basis for differenti- 
ating managers from nonmanagers. 

It is often pointed out that managers must represent the enterprise, or 
some portion thereof, in dealings with such external units as stockholders, 
consumers, suppliers of goods used by the enterprise, organized labor, 
competitors (either individually or in trade associations), government 
units, the general public, etc. Managers speak and act for the units they 
manage. They often enter into contracts with an external unit, acting as 

6 Many writers use the word "control" to stand for that which I designate by 
"supervision." Others follow the practice I do. Still others use the words "appraisal" 
or "evaluation." Here, as before, no available word is completely satisfactory to 
connote all that one would desire. The weakness of "supervision" is that it often 
implies work carried on by those near the bottom of the managerial hierarchy. Re- 
gardless of the word used, the ideas behind the words are usually similar; e.g., 
Brech: " 'Continuous control and supervision of the activities of the organisation' is 
nothing else than the obverse of planning. It is the task of seeing that the plans laid 
down are being currently and effectively carried out, or establishing good reasons for 
failures and departures." (E. F. L. Brech, The Nature and Significance of Manage- 
ment [London, 1946I, p. 16). Fayol: "Control ... he [Fayol] regarded as an aspect 
of Administration. He denned it as, 'Seeing that everything is being carried out in ac- 
cordance with the plan which has been adopted, the orders which have been given, 
and the principles which have been laid down.' " (L. Urwick, The Elements of Ad- 
ministration [New York: Harper & Brothers, 1943], p. 105.) Copeland: "After a plan 
is in operation, the executive has the task of checking up to learn whether it is 
being carried out in accordance with the policy formulated to meet the conditions 
and with the sequence and timing decided upon. This follow-through is the essence 
of executive control" (Melvin T. Copeland, "The Job of an Executive," Harvard 
Business Review, XVIII, No. 2 [winter, 1940], 158). 

The Manager Concept 2 5 

an agent of the enterprise of which they are a member. All of this is true, 
but it is also true, at times, of nonmanagers as well. The act of representa- 
tion on the part of a nonmanager is never sufficient in practice to give him 
the status of a manager. 

In this connection, it is important to note that all the services con- 
tributed by certain individuals to an enterprise are not necessarily 
managerial in character. Managers often reserve to themselves some non- 
managerial work to perform which they consider too important to dele- 
gate to someone else. Much representation work performed by managers 
is of this character. 

It was also noted that many writers consider co-ordination to be a 
function of managers. Again the writer disagrees with such a point of 


At the outset of this article, it was indicated that primary concern 
would be with differentiating managerial services from nonmanagerial 
services and that this could best be accomplished by isolating those func- 
tions performed exclusively by managers. Such isolation has been at- 
tempted. Now the separate threads of this article can be drawn together 
and combined into meaningful conclusions. 

It is the thesis of this study that managers are those who use formal 
authority to organize, direct, or control responsible subordinates {and 
therefore, indirectly, the groups or complexes which they may head) in 
order that all service contributions be co-ordinated in the attainment of an 
enterprise purpose. 

Managers always stand in a relationship of formal authority over 
subordinates who, in turn, are responsible to their superior. Managers use 
formal authority in order to execute the functions of managers— organiza- 
tion, direction, and control. The objective of the execution of the func- 
tions is the co-ordination of service contributions in the attainment of an 
enterprise purpose. An individual is not a manager, does not manage, un- 
less he has and uses formal authority to organize, direct, or control re- 
sponsible subordinates. In business enterprises individuals may be called 
managers who do not fit this specification. From our point of view, they 
are managers in name only. In determining who is a manager, one must 
look to functions performed and not to titles. Unless he conforms to this 
specification, he is a nonmanager. 

When specialized services are being contributed toward the attainment 
of an enterprise purpose, co-ordination is essential. This co-ordination is 
supplied by managers through their execution of the functions of man- 
agers. Many writers consider co-ordination to be a function of managers. 

2 6 What Is Control? 

Co-ordination is not properly a function; it is something to be achieved. 
And it is achieved by adequate organization, direction, and control. The 
services of managers (involving organization, direction, and control) are 
necessary to co-ordinate the specialized service-contributions of the units 
which they head in the attainment of an enterprise purpose. The services 
of managers are needed for no other reason. 

From what has been said it can be seen that managers can be differenti- 
ated from nonmanagers. 7 Managers always have subordinates; nonman- 
agers never do. Nonmanagers may organize, direct, or control (in a sense) 
themselves or the material objects with which they work, but they never 
organize, direct, or control responsible subordinates. Furthermore, it is 
important to see that managers as well as nonmanagers may be managed. 
All managers, except the one (or ones) who heads the supreme complex, 
are managed by their superiors. They, in turn, manage their subordinates. 
The crucial distinction to be made is between managers and nonmanagers 
—not between managers and the managed. 

One final point needs emphasis. The services of all individuals are essen- 
tial if the purpose of the enterprise is to be attained. The .distinction be- 
tween managers and nonmanagers is based on differences in the types of 
specialized services contributed by the two groups and on no other cri- 


Alex W. Rathe* 

Professor Alex W. Rathe regards management control as an executive's 
instrument panel. An important but brief element in this article is the 
"feedback" of current control data to the planning group for use in future 
planning; this is illustrated also in the diagram of the author's concept of 
management control. 

7 This statement is contrary to the opinion of some other writers. For example, 
Brown says: "Management has been very widely used in the literature of business to 
denote both the act of management and those who manage. As so used, it represents 
the concept of an upper tier of endeavor which can never be exactly defined. Be- 
cause it cannot be defined, this treatise has concluded that, however useful the notion 
may be in treating of administration in general, it does not help in understanding 
organization. Those who 'manage' have responsibilities which differ from other re- 
sponsibilities only in scope. Their acts are acts of administration differing only in 
scope from any act of administration. The authority which they exercise differs only 
in like manner. Indeed, the distinction between management and labor is a highly 
artificial one which probably has not helped our social economy." (Alvin Brown, 
Organization [New York: Hibbert Printing Co., 1945], p. 104, n. 1.) 

*From Advanced Management, XV, 3 (1950), 9-1 1. Reprinted by permission of 
Advanced Management. 

Management Control 2 7 

No pilot flies a plane without navigating instruments which guide him 
to his target. The Executive's Instrument Panel which guides progressive 
business to its target is Management Control. 


"Control" of operations is good only when it helps operations. Does 
this not mean . . . 

1. That control must give warnings of obstacles which lie ahead and 
that it should map out plans to avoid them. 

2. That control must examine past and present activities so as to search 
out weaknesses which can be eliminated in the future, and that it has to 
ascertain the results of its planning. 

The first point covers planning activities. They determine what should 
take place. The second indicates control activities. They determine what 
did take place. Separated, each is vulnerable. Grouped together, planning 
and control work is a powerful managerial instrument. It operates in a 
cycle. Plans are formulated, tried out, and checked; their results influence 
new plans which are put into operation and tested again; and so forth in a 
continuing chain reaction. 

When properly organized, these techniques produce such potent results 
that they have been called "the new look in management." In less com- 
mercialized appellation, the interplay of planning and control has become 
known as Management Control. 

We need both planning and control in this task of improving manage- 
ment. The primary reason is perhaps that control is most effectively ex- 
ercised before something goes off the beam; for that, you need to know 
what should take place— what the plans are— so as to be able to guide mat- 
ters there. 

Furthermore, it is difficult to separate planning and control in many 
instances. The budget, for example, is a planning tool; it sets forth the 
schedule of operations decided upon. But marking against the columns of 
"budgeted" data the corresponding figures of "actual" performance brings 
forth the control aspects of the budget; this is particularly so when there 
are added comparable data on operations during the previous season or on 
activities in some other place. 

Time standards serve both planning and control because they facilitate 
planning while comparison of actual with standard time serves the con- 
trol end. 


The planning phase of management control determines how operations 
should proceed. This is a formidable job; depending upon individual cir- 
cumstances, it varies considerably in scope; Figure 1 shows a relatively 

28 What Is Control? 

wide range. So many of the planning activities are problems which the 
line executive, in the pressure of everyday operating responsibilities, is 
forced to postpone for the proverbial tomorrow which too often never 
comes. He usually welcomes the assistance in these projects of a staff 
agency such as the Controller's Division which will be sketched here. 

Planning has two segments. First, the operating program must be de- 
signed and secondly, managerial tools are to be developed which permit 
this program to be executed in the best manner. 

With policies as reference points, long range plans and short term pro- 
grams are designed. Long range plans peer three, five, or ten years into 
the future. They cover any and all phases of corporate activities. 

For the short range program the foundation is laid by market analysis 
which ascertains the market potential. Sales forecasts calculate the prob- 
able share of this total demand which the company can expect to fill. This 
estimate is reconciled with the results of a separate study of economic 
conditions and then, the corporation's over-all short term program evolves, 
with its component engineering, manufacturing, marketing, and other 
divisional programs. 


The first control task is to locate those spots in all parts of corporate 
activities which compile significant factual data on performance and to 
synchronize them into one over-all management control network. 

Financial data come from three main reservoirs. Cost accounting is the 
custodian of the expense records of operations, parts, products, depart- 
ments, or other subdivisions. The second carrier of financial information is 
the budget. Its clear-cut schedules, listing the plans for all important fi- 
nancial transactions, represent the operating program in monetary terms. 
Financial auditing is the third of the contributories to the stream of finan- 
cial data. 

All other data, trends, events, and additional facts should be collected, 
recorded, and analyzed by a statistical group; its pipelines connect with 
production control, payroll, timekeeping, purchasing, quality control, 
sales, and many other records. 

After all these control data are collected, they are analyzed. 

The really responsible part of the job is the evaluation of these findings. 
This is where the management control group meets its acid test. Data and 
facts must be interpreted in the light of existing circumstances. They 
should be evaluated from the point of view of over-all management. And 
most importantly, they have to be examined in their relation to, and in- 
fluence upon, one another. Thorough consideration has to be given to 
economic, human and public relations as well as many other aspects. Main 

Management Control 20 

emphasis does not rest so much on analysis but on synthesis of the com- 
ponents of the findings. 

The results of the evaluation are usually compiled in reports. They are 
also often presented to the executives concerned in less formal fashion, 
somewhat in the manner in which a physician would discuss with a col- 
league the results of his diagnosis and experiments. Whether it is the 
counsel of the doctor or of the controller, it is entirely within the in- 
dividual's discretion to accept or reject the advice offered. 

The findings are finally channeled back to the planning group so that 
they will be considered in future developments, policies, plans, and pro- 
grams. This "feed-back" completes the cycle of management control 

The bird's-eye view shows at once that management control is not 
synonymous with economics, statistics, accounting, or engineering. It 
contains elements of all of these, and much more because, as the Executive's 
Instrument Panel, it gives an account of all operational aspects in all 
phases of the company. 


The first requirement for organizing the management control function 
will always remain that the solution is acceptable on the psychological 

That is why management control is usually decentralized. In this manner, 
closer relationships become possible and greater interest in the work is 
generated because everyone sees what he contributes and what his con- 
tributions mean in the larger picture. The nearer you get to the specific 
operation, the more effective is your control and the more readily are 
control findings accepted. 

Decentralized, each division, store, or plant has its own management 
control staff, its own subsidiary instrument panel. Here are performed all 
planning and control activities for the group. A selection of typical assign- 
ments for the functional divisions of a manufacturing company might 
for instance be . 


Marketing sales forecast 

Manufacturing manufacturing program 

delivery timetable 
production control schedule 


Finance accounting procedures 

Production methods 


What Is Control? 




development cost 
divisional budget 
other engineering data 

morale survey 

Management Control: The Executive's Instrument Panel 




I , 




















This then leaves the Controller with four main spheres: 
First, he has to connect the subsidiary instruments in the various divisions 
into an over-all corporate instrument panel from which top executives and 
division managers alike obtain their reading on the course and the per- 
formance of the enterprise as a whole. 

Second, he handles projects of mutual concern and acts as a clearing 
house for the exchange of data, information, and advice on problems of 

Management Control 3 1 

common interest to several divisions (such as corporate organization 

Third, he is responsible for corporation-wide problems which affect all 
divisions (such as economic studies, over-all policies, etc.). 

Fourth, his staff is available to all other divisions as the company's own 
management consultant whenever such services are desired; and this is 
usually the case with a great variety of problems. 


Three considerations are paramount if management control is to be 

One, planning and control activities must be recognized as components 
of one task, that of managing; they are not a tool but a part of manage- 

Two, the profitable functioning of management control depends on 
reliable interconnection not merely of some but of all planning and con- 
trol work wherever it may be carried on in a firm. 

Three, decision by factual data is to be emphasized, rather than a hit-or- 
miss approach, because you can't get the better way of tomorrow until 
you have all the facts of today. 

Management control reinforces the executive whom it serves. It multi- 
plies his effectiveness. But it has no command authority. It is confined to 
the authority of ideas, findings, and suggestions. It does not give orders. 
It must sell its recommendations. Or better still, it must present its thoughts 
in such a form that the operating executives want to buy. 

Management control backstops management. In its planning phase, it 
determines the atmosphere— economic and otherwise— within which opera- 
tions are expected to be carried out. It lays plans. It sets targets. It suggests 
the best road map and timetable through which the executive can reach 
his target. 

And then, in its control phase, management control tells him what kind 
of a job he really did. It helps uncover errors made while pursuing the 
plans or deviating from them. It gives this information in clear-cut facts 
and figures. There is no hunch, no guesswork, no opinions. Management 
control is fact not man control. 

In addition to supplying factual information on all phases of operations, 
management control evaluates these facts. It gauges actual performance 
against plans. It interprets results. It adds recommendations on how to pre- 
vent weak spots and avoid pitfalls in the future. Finally, it supplies practical 
ideas on how to proceed to make an ever-better record of profit and 
service and it channels them back into the planning phase. 

- 2 What Is Control? 


Bradford Cadmus* 

The need for the periodic and comprehensive survey and appraisal of 
all functions of a unit or cross-section of an enterprise is gaining increasing 
recognition. Who is to do this job? One answer is the internal auditor, and 
Mr. Cadmus gives a program approach for the guidance of management as 
well as the internal auditor. 

We live in an age in which some previously clear-cut distinctions have 
broken down or have disappeared. For example, when I took a course in 
chemistry, there was a definite division between organic and inorganic 
chemistry. The chemist of today gives little or no recognition to this 
separation. Similarly, matter and energy were considered to be separate 
until atomic scientists changed that idea. 

In the same way, financial and operational auditing are not-and should 
not be-separate and distinct types of auditing. To a very considerable 
degree, the techniques that we propose to describe as applicable to opera- 
tional auditing apply to internal auditing of any description and in any 
department of a company. Yet there are certain differences that must be 
taken into account in the auditor's approach and his work in various de- 
partments. From the auditor's standpoint, these departments may be 
divided into two groups— corporate service and operating. 

By corporate service departments we mean departments handling such 
functions as accounting, treasury, legal, taxes and insurance. To a very 
large degree, the work of these departments is conditioned by the require- 
ments and the reports of the rest of the business. What will be recorded 
and reported is based upon decisions, policies and happenings that have 
originated in other departments. In audits of corporate service departments 
the internal auditor makes a constructive contribution as (1) he makes 
findings in matters of policy, procedure and reporting which lead to more 
effective operation of these departments and (2) as he finds clues in these 
audits which lead back to the operating departments. 

Now, I propose that we center our attention on all corporate functions 
other than those which have just been defined as corporate service. To put 
it positively, let us assume that you are faced with an assignment from 
management to conduct audits of the purchasing, traffic, production, sales, 
advertising, engineering and research operations of your company. 

*From The Internal Auditor, XVII, 1 (i960), 28-39. Reprinted by permission of 
the author and the Institute of Internal Auditors. 

Operational Auditing 

The first reaction of one who is not an internal auditor-and possibly of 
some internal auditors who have concentrated on financial audits-might be 
to wonder what sort of superman would have the knowledge and back- 
ground to cover these diverse operations and be able to make a construc- 
tive contribution. The answer is simple-the internal auditor is an expert 
in control, not in operations. Every department of a company has organ- 
ization, procedures, records, reports and some sort of formal or informal 
standards by which it appraises its performance. These are the controls, 
and these controls will be found in some form or other in every operating 
department. So the internal auditor can apply his knowledge of the tech- 
niques and principles of control to any department, just as a man familiar 
with the techniques of purchasing uses his talents in the purchasing de- 
partment to buy what is required throughout the company. 


Description of the « elements of control will be found in a number of 
publications. To give the essentials: 

i. Organization control requires that each employee know his place in 
the organization and exactly what authority and responsibility have been 
assigned to him. Additionally, organizational control requires adequate 
checks and balances with separation between operating responsibility and 
the accounting for that responsibility. 

2. Procedures have the objective of dividing and defining the work to 
be done into logical, understandable- sections that specify the work and 
responsibility of each employee. 

3. Records include accounting and all other records which show what 
has occurred— as a basis for information and reports. 

4. Reports are a major means of management control. Both records and 
reports must be prompt, accurate, concise and complete. Reports must be 
(1) impartial in presenting a fair picture and (2) adaptable to administra- 
tive use, following the pattern of organizational responsibility. 

5. Standards of performance provide the means of judging how what 
has occurred compares with what was expected or planned. Commonly 
used standards of performance are budgets, standard costs and comparisons 
with preceding periods. 


Before we discuss the actual work of audit in the operating department, 
let us summarize the talents that the internal auditor brings to his task that 
qualify him to do a profitable and constructive piece of work in his 

What Is Control? 

i. His knowledge of the techniques and philosophy of control, which 

have just been described. 

2. His knowledge of his company. In this respect, he will usually be 
better informed than those with whom he is dealing. Through actual ex- 
perience in various departments, he will know what is done, how it is 
done and how it fits into the over-all company operations. 

3. His business sense-that sense of proportion which weighs the relative 
importance of each element of a situation in relation to the welfare of 
the business as a whole. 

4. His innate curiosity about the "what," "why," "who" and "how" of 
each operation. The "why" is particularly important-and the auditor must 
be patient and persistent until he is completely satisfied with the answer. 


Let us assume that the internal auditor is making an audit of an operat- 
ing department for the first time. First he must sit down with the depart- 
ment head and explain the purpose of his audit. He explains that he is 
not there to pose as an expert in the function of that department. He is 
there to examine, appraise and report on the controls which govern the 
work of the department. He is there to learn of any apparent deficiencies 
or failures of control or other relationships between that department and 
the other departments of the company. (In this area, he is almost certain to 
arouse interest, since failings of this sort are very common.) 

The next step is to acquire general familiarity with the basic operations 
and objectives of the department-since he must know in general what the 
job of the department is in order to examine and appraise the effectiveness 
of the control structure which is set up to fulfill that responsibility. 

Some operating department heads like to throw an aura of mystery 
about their work, implying that one can begin to understand their work 
and problems only after years of specific experience. Very often such indi- 
viduals like to talk-and from their talking, and with a few judicious ques- 
tions, the auditor can develop enough information to provide the essential 
background. The auditor must explain, of course, that he is not concerned 
with becoming an operating expert-he merely wants a general picture of 
policies, procedures and problems. The executive who gives a straight- 
forward description is fortunately more common, and the auditor need do 
little but listen and guide the conversation with an occasional question. 

The important objective in this familiarization is to learn how the opera- 
tions are viewed by the man who runs them. Even when the auditor has 
a fair knowledge of the operations himself, he should not volunteer it- 
since this will defeat the objective of learning from the operating executive. 

Operational Auditing 3 5 

In these conversations with the department head, it is important to 
learn the means by which that executive decides how well the department 
has performed— both in its individual units and as a whole. Some interest- 
ing and occasionally surprising answers will be given when the question is 
asked "How do you decide how good a job John Brown is doing?" In this 
particular question, the auditor is trying to develop the standards of per- 
formances that the department applies to its work. 


The general plan of making an operational audit has been well described 
in the Research Committee Report on "Internal Audit and Control of a 
Traffic Department." Rather than to paraphrase, I quote directly from the 


Viewed in its broad aspects, the approach and general plan of an audit of an 
operating department is the same as an audit of a financial department. The 
steps by the internal auditor may be summarized as : ( 1 ) familiarization, 
(2) verification, (3) evaluation and (4) reporting. 

Familiarization. In the first step the internal auditor, through discussions 
with departmental personnel, acquaints himself with the operating objectives 
and problems of the department. He then proceeds to learn how these ob- 
jectives and problems are met and controlled by departmental management. 

First comes the organization structure. Here the auditor is interested in 
learning the assignments of authority and responsibility and the interrelation- 
ships with other departments and between subordinate groups within the 

Next comes the review of the procedures which govern the work of the de- 
partment. The concern here is to learn the degree to which procedures are 
definite and complete, so that no "twilight" zones will lead to oversight or 
conflict of responsibility. 

Written procedures are a necessity in practically all operations. Through 
them, individual employees know what they have to do— and those directly 
responsible and others are informed as to the scope of responsibility and the 
methods of operation. Written procedures establish standards of performance. 
When a violation occurs, it provides a definite signal that further investigation 
should be made. Something may be unusual about a particular transaction, or 
the procedure itself may be incomplete or incorrect. 

Verification. Having familiarized himself with the general responsibilities of 
the operating department and the manner in which the department controls 
its operations, the internal auditor then proceeds to verification. In this portion 
of his work, the objective is to learn whether the actual operations and assign- 
ment of responsibility follow the plans prescribed by departmental manage- 

Verification requires that a selected sample of transactions or a selected area 
of work be examined in detail. The original size of the sample or the area will 


What Is Control? 

depend on the judgment of the auditor and will usually be comparatively 
small. If the original tests indicate that further sampling is desirable to reach a 
conclusion, then such sampling follows: 

In verification, the objective of the internal auditor is fact-finding. He is 
concerned with learning: 

i. Do the organization structure and assignment of responsibility follow the 
control plans of departmental management? 

2. Are procedures prescribed by management being followed? 

3. Are the internal controls established by departmental management being 

4. How are the internal controls co-ordinated with other operating depart- 

Evaluation. Having determined what the plans are and how effectively they 
are being followed, the internal auditor then proceeds to the work of evalua- 
tion and appraisal. In this phase, his problems will be: 

1. Do departmental controls conform with company policy? If not, what 
appears to be the reason? 

2. If there are numerous deviations from established procedure, does the 
reason lie with the procedures or with other factors? 

3. Do departmental controls automatically reveal to management the out-of- 
the-ordinary situation that should have attention? 

4. Do departmental controls appear effective in operation? 

Reporting. In financial areas, a definite opinion can usually be given as to 
effectiveness of controls— since ineffective controls will lead to questionable 
operations or incorrect results. 

In operating departments, such opinions are often not possible. To quote 
from one auditor: 

"Operational audits have been most successful when the auditor was not in 
the position of matching his opinion as to the effectiveness of a control or the 
need for an additional control with the various levels of management responsi- 
ble for such control. Normally such a position can be avoided by providing, 
through normal auditing techniques, satisfactory evidence that a control is not 
effective and should be corrected or that an uncontrolled area should be 
placed under control. But it is absolutely essential that evidence be developed 
and produced in support of the auditor's position." 

Following this thought, it becomes evident that the report is primarily a 
pointing out to departmental and general management how controls are oper- 
ating—in theory and in practice, both within the operating department and in 
relation to other departments. 

An essential of reporting is that the report be discussed and reviewed in 
complete detail with the operating department before submission in final form. 
This review will insure that there is agreement in factual matters and gives 
the operating department a chance to suggest the course of corrective action 
that may be indicated by the facts of the report. 


Because the internal auditor's work in operating departments requires 
that he become familiar with the problems and the operations of these de- 
partments, there is apt to be some confusion as to the responsibility of the 

Operational Auditing ^ _ 

internal auditor and of those engaged in organization and methods or 
operations research work. In actuality, there is no reason for confusion or 

Organization and methods and operations research personnel study the 
work of an operating department with an objective of developing pro- 
cedures, reports and other routines and controls to govern the day-to-day 
work of the department. The internal auditor then appraises the operation 
of the controls which these other groups have established. He will point 
out strengths and weaknesses-but he does not establish the controls. That 
is not his job. This separation of responsibility is expressed in the State- 
ment of Responsibilities of the Internal Auditor-"internal auditors should 
not develop procedures, prepare records, or engage in any other activity 
which they normally would be expected to review and appraise." 


I wish it were possible to describe some sort of sales campaign by which 
management could be convinced of the benefits of operational auditing. 
Unfortunately, this cannot be done, since managements and auditing de- 
partments are composed of individuals-with individual talents, ideas and 
preferences. For this reason, the selling job is an individual one in every 

As far as the internal auditor is concerned, I can offer certain sug- 

i. The internal auditor should prepare himself, through study of his 
company and of internal auditing material, to do the broad job that is 
required in operational auditing. 

2. When opportunity offers, the internal auditor should bring to 
management's attention what other companies have done in the operational 
auditing field. Institute publications and similar material are one way of 
doing this. 

3. The internal auditor should do a broad and outstanding job in the 
field to which he is presently assigned. Then he is apt to find that his 
field of effort will be broadened. This approach is summed up in a little 
saying that my mother was fond of, "The reward of work well done is 
more work." 

In talking of mangement, I mean executive management and not the 
managers of operating departments. It is essential that the auditor come 
into operating departments with the wholehearted backing of executive 
management. If he does not have this, he had better not start. Once in the 
departments, he should be able to demonstrate to both departmental and 
executive management that he can perform a constructive service. This 

Q What Is Control? 

usually is the case-I have no personal knowledge of any situation where 
there has been a retreat from an operational auditing program once it has 
been established. 


The objectives and scope of internal auditing are succinctly outlined 
in the following statement: 


Internal auditing is an independent appraisal activity within an organ- 
ization for the review of accounting, financial and other operations as a 
basis for service to management. It is a managerial control, which functions 
by measuring and evaluating the effectiveness of other controls. 


The over-all objective of internal auditing is to assist all members of 
management in the effective discharge of their responsibilities, by furnish- 
ing them with objective analyses, appraisals, recommendations and perti- 
nent comments concerning the activities reviewed. The internal auditor 
therefore should be concerned with any phase of business activity wherein 
he can be of service to management. The attainment of this over-all 
objective of service to management should involve such activities as: 

Reviewing and appraising the soundness, adequacy and application of 

accounting, financial and operating controls. 

Ascertaining the extent of compliance with established policies, plans 

and procedures. 

Ascertaining the extent to which company assets are accounted for, and 

safeguarded from losses of all kinds. 

Ascertaining the reliability of accounting and other data developed 

within the organization. 

Appraising the quality of performance in carrying out assigned respon- 

* From the Institute of Internal Auditors, 1957. Reprinted by permission. 

What Is a Controller? ^q 


Internal auditing is a staff function rather than a line function. There- 
fore the internal auditor does not exercise direct authority over other 
persons in the organization, whose work he reviews. 

The internal auditor should be free to review and appraise policies, plans, 
procedures, and records; but his review and appraisal does not in any way 
relieve other persons in the organization of the responsibilities assigned 
to them. 


Independence is essential to the effectiveness of the internal auditing 
program. This independence has two major aspects. 

i. The organizational status of the internal auditor and the support 
accorded to him by management are major determinants of the range and 
value of the services which management will obtain from the internal 
auditing function. The head of the internal auditing department, therefore, 
should be responsible to an officer of sufficient rank in the organization as 
will assure a broad scope of activities, and adequate consideration of and 
effective action on the findings or recommendations made by him. 

2. Since complete objectivity is essential to the audit function, internal 
auditors should not develop and install procedures, prepare records, or 
engage in any other activity which they normally would be expected to 
review and appraise. 


The duties and responsibilities of a controller are listed in the following 
statement issued by the Controllers Institute of America. 

The institute's concept of the function of controllership is: 
i. To establish, co-ordinate and administer, as an integral part of man- 
agement, an adequate plan for the control of operations. Such a plan 
would provide, to the extent required in the business, profit planning, 
programs for capital investing and for financing, sales forecasts, expense 

* From Controllers Institute of America, 1959, 12 pp. Reprinted by permission of 
the Controllers Institute of America. 

What Is Control? 

budgets and cost standards, together with the necessary procedures to ef- 
fectuate the plan. 

2. To compare performance with operating plans and standards, and 
to report and interpret the results of operations to all levels of management 
and to the owners of the business. This function includes the formulation 
and administration of accounting policy and the compilation of statistical 
records and special reports as required. 

3. To consult with all segments of management responsible for policy 
or action concerning any phase of the operation of the business as it relates 
to the attainment of objectives and the effectiveness of policies, organ- 
ization structure and procedures. 

4. To administer tax policies and procedures. 

5. To supervise or co-ordinate the preparation of reports to govern- 
mental agencies. 

6. To assure fiscal protection for the assets of the business through 
adequate internal control and proper insurance coverage. 

7. To continuously appraise economic and social forces, and govern- 
ment influences, and interpret their effect upon the business. 


A special report of the Committee on Auditing Procedures of the 
American Institute of Accountants (now known as the American Institute 
of Certified Public Accountants) discusses the elements of a co-ordinated 
system and its importance to management and the independent public 
accountant. Included in the report is a definition of internal control, the 
essentials of which follow. 

Internal control comprises the plan of organization and all of the co- 
ordinate methods and measures adopted within a business to safeguard its 
assets, check the accuracy and reliability of its accounting data, promote 
operational efficiency, and encourage adherence to prescribed managerial 
policies. This definition possibly is broader than the meaning sometimes 
attributed to the term. It recognizes that a "system" of internal control 
extends beyond those matters which relate directly to the functions of the 
accounting and financial departments. Such a system might include budget- 
ary control, standard costs, periodic operating reports, statistical analyses 

* Reprinted with permission from Internal Control, published in 1949 by the 
American Institute of Certified Public Accountants. 

Internal Control 

4 1 

and the dissemination thereof, a training program designed to aid personnel 
in meeting their responsibilities, and an internal audit staff to provide 
additional assurance to management as to the adequacy of its outlined 
procedures and the extent to which they are being effectively carried out. 
It properly comprehends activities in other fields as, for example, time and 
motion studies which are of an engineering nature, and use of quality 
controls through a system of inspection which fundamentally is a pro- 
duction function. 

What may be said to be the characteristics of a satisfactory system of 
internal control? Certainly, they would include 

A plan of organization which provides appropriate segregation of 

functional responsibilities, 

A system of authorization and record procedures adequate to provide 

reasonable accounting control over assets, liabilities, revenues and 


Sound practices to be followed in performance of duties and functions 

of each of the organizational departments, and 

A degree of quality of personnel commensurate with responsibilities. 

These elements, as important as each is in its own right, are all so 
basic to proper internal control that serious deficiencies in any one nor- 
mally would preclude successful operation of the system. For example, no 
plan of authorization and record procedures for accounting control may 
be considered adequate without personnel capable of performing the 
procedures designed to make such a system work, nor can one consider 
the practices followed in the performance of duties in the organizational 
departments sound unless there is departmental independence so that re- 
sonsibilities can be placed and interdepartmental controls enforced. 


An important criterion as to the adequacy of any plan of organization 
is the extent to which it provides for organizational independence as 
between operating, custodian, and accounting (including internal audit- 
ing) departments. Organizational independence does not imply the erection 
of any barriers preventing frequent consultation between departments to 
devise means of smoothing the flow of work and increasing the over-all 
efficiency of operation. The work of all departments must be closely inte- 
grated and co-ordinated and, to that end, co-operation is essential. The basis 
for the separation rests on the premise that no department should control 
the accounting records relating to its own operations. It represents an 
extension at the departmental level of the oft-repeated principle that no 

What Is Control? 
4 2 

one person should control all phases of a transaction without the inter- 
vention of some other person or persons who afford a cross-check. With- 
out such a separation, the accounting records may be so manipulated as to 
make the detection of errors and fraud exteremely difficult, if not im- 



Arnold F. Emch* 

The author asserts that the functions of planning and performance eval- 
uation are not independent of organization. He relates controls to the 
specific responsibilities and activities of each executive position, which he 
calls functional control, thus making it an integral part of each executive's 
duties and not something which is done by a "controller." Control, thus 
envisaged, emphasizes the follow-through activities which result from an 
awareness of what is expected and what is accomplished. 

In today's competitive economy there is a tremendous premium on 
initiative in management. Although intelligent policy making and plan- 
ning are, as always, of decisive importance, top-level planners feel more 
dependent than ever on those "centers of initiative" down the line. 

How does the concept of "control" fit into this picture, particularly 
with respect to the problem of initiative on the part of executives who 
are charged with the responsibility of getting things done? Is control a 
boon or a barrier to initiative? This question is widening a four- way split 
of executive opinion in industry as well as in governmental and nonprofit 
enterprise. One group is unqualifiedly for something or other they call 
"control"; a second group looks on the whole business with a jaundiced 

* From Harvard Business Review, XXXII, 4 (1954), 92-98. Reprinted by permission 
of the Harvard Business Review. 


Control and the Administrator 

eye; a third group finds itself uncomfortably in the middle with the feel- 
ing that there is something wrong with each of the extreme points of 
view; and a fourth group subscribes to the functional concept that will be 
presented in the following pages. 


The bewildering difference of opinion and practice in this area makes 
no sense at all until we remember that control-perhaps more than any 
other major management function-reflects the personalities and attitudes 
of those at the top. Let us see what some of these executives look like. 


At one extreme is the Captain Queeg type of Caine Mutiny fame, who 
insists on the letter of the law to such an extent that war can rage all 
around him while he is trying to find out who ate the quart of straw- 
berries. From the standpoint of management, what was the matter with 
Captain Queeg? Two things-one psychological, the other methodological. 

As Herman Wouk, the author of the book, painted him, Captain Queeg 
was fundamentally so insecure and suspicious that he had to know every 
last detail of what every last tar aboard the U.S.S. Caine was doing in 
order to protect himself and to assure himself that all policies, rules, and 
regulations of the Navy were observed down to the last minutia. What he 
forgot in this preoccupation with detail was (a) that he and his kind 
might lose the war, and (b) that he was stifling all initiative, interest, and 
enthusiasm on the part of his men. In short, he was, through his interpreta- 
tion of control, forgetting the strategic objectives of the war and, at the 
same time, creating a serious morale problem in his crew. 

The situation on Captain Queeg's ship was a typical example of the 
stultifying effect of rigidly confining strait-jacket "control procedures" 
on the imagination and intellect of men. In the words of Keef er, one of the 
principals of the story, it was a situation in which "the work has been 
fragmentized by a few excellent brains at the top, on the assumption that 
near-morons will be responsible for each fragment." 

In business this type of control does not of course bring about mutiny 
in the usual sense. What it does instead is to create (a) an undertone of 
frustration and a sense of futility; (b) a pattern of alibis, truth-slantings, 
and downright dishonesty; (c) outright conflicts among executives, due 
to exercise of control by some who have no corresponding responsibility 
for action; (d) a gradual diminution of the use of initiative and judgment 
to the point where executives are absorbed in clerical detail and meticu- 
lous line-by-line paper-pushing practices; and (e) an increase in executive 

Control Means Action 45 

turnover as a result of the more intelligent and courageous personnel look- 
ing elsewhere for a more favorable condition for the exercise of their 
real talents. 

I have cited a fictitious character as an example; but actually, in business, 
government, and nonprofit enterprises, there are literally hundreds of 
Captain Queeg's counterparts. These men build up a system of control 
that in the end defeats the purposes for which it was originally intended. 
The paper, the personnel, the money, and the time expended on forms, 
reports, manuals, bulletins, and statistics in support of these misbegotten 
control systems are enough to stagger the most eager imagination. As 
pointed out in the ACME Reporter (bulletin of the Association of Con- 
sulting Management Engineers) : 

It often develops to the point where mere following of procedures becomes 
more important than carrying out policies and striving toward objectives. 
The result is a creeping, self -propagating bureaucracy. 

Do these burdensome control systems ever accomplish their original 
purpose? Do they not have some merit? All I can say is that in my experi- 
ence as a management consultant to different kinds of businesses over the 
past 20 years, I have never seen such a setup or such an operation pay off. 

Quite the contrary! I have found all the shortcomings of the usual dic- 
tatorial and all-embracing regime, including the inevitably weak, insecure, 
or overambitious executive at the top; mountains of forms, reports, man- 
uals, directives, and interpretative bulletins which nobody reads or which, 
if they are read, require special staffs to pore over them to keep executives 
abreast; bad morale in the second and third layers of management; and, 
believe it or not, usually poor planning and lack of significant management 
information when and where it is needed— the very things you would think 
could be accomplished by these immense, involved, and demanding sys- 
tems of control. 


Now let us look at Captain Queeg's opposite, the executive who is 
pretty sure of himself and who inclines to be a "good egg" with demo- 
cratic impulses. He does not like control and he does not want any part of 
it. It is repugnant to him. His philosophy of management is to shove his 
men off the dock and make them swim. That is the way they can learn 
best. "Sure, they'll make mistakes," he says with a confident, genial grin; 
"that's good for them; that's the way they'll learn the facts of life." His is 
the freedom-of-enterprise point of view in the extreme. 

But despite his rodeo-bronco-swim-for-your-life approach to manage- 
ment, this Will Rogers sooner or later will want to know where his busi- 
ness is going and how well it is doing on its way. Instead of having a "con- 
troller" in any usual sense— this of course he would not tolerate— he sets out 

46 Control and the Administrator 

to find himself a bright, young, likable, willing fellow, and without much 
ado appoints him "assistant to the president." 

There are all kinds of things that Mr. Rogers will want to know— usually 
pronto— and no ready information will be available to supply this need. So 
he will call in his assistant, throw him the ball without concern or pre- 
meditation, and let it go at that. 

It takes no genius to figure out what this apparently simple practice will 
lead to. The assistant is eager, able, and willing; and he goes about these re- 
peated assignments with energy and dispatch. It does not take long for 
him to realize that he does not have certain information at hand. He begins 
to build up a little system all his own. His desk drawers and filing cases 
fill up with special data. He takes an increasing interest in the budget and 
begins to question department heads as to the justification of certain items. 
He even has misgivings on just how he ought to go about getting some of 
the information that he knows he will eventually need. 

But the rest of the staff soon comes to realize that, although the new- 
comer is only the assistant to the president whereas they are vice presi- 
dents or managers, this young fellow has the ear of the boss and is begin- 
ning to analyze things and increasingly make judgments for the president. 
Perhaps they had better play ball with the young man, they think, if they 
are to stay in the good graces of Mr. Rogers. And so it goes. A roundabout 
way of control is devised— never direct, never through channels, never 
through organization objectives, or policies, or clear-cut statements of 
basic procedure. Nobody quite knows any more what he should or should 
not do, but everyone is quite certain that "cooperation" with the assistant 
is the order of the day. 


Now what is the matter with all this? What is the matter with Captain 
Queeg's or Mr. Rogers' approach to control? It boils down to this: Each 
approach shows a complete lack of understanding or appreciation of what 
control is trying to accomplish in an enterprise, and of how to get execu- 
tives to assume and carry out effectively their proper responsibilities. Put 
very simply, there can be too much control and there can be too little con- 
trol; in both cases there is a misunderstanding or a corruption of control 
in the necessary and sound sense. 

Let us look at it in another way. What is the problem we are dealing 
with when we talk of "control"? Actually it is simple: We have a job 
to do— a line of services or products to make and sell at a profit. There are 
a number of persons involved in the doing of that job. Hence we organize 
ourselves in some fashion so that each one of us has specific, assigned tasks, 
all more or less related to one another. And we try to see that each key 

Control Means A ction 47 

individual has a clear understanding of his functions, of his lines of 
authority downward, and of his line of responsibility upward. 

But if we should go only this far, we would not go far enough. We 
must also determine what each of these individuals needs in the way of 
facts and figures in order to perform his job effectively. This, then, is 
the problem of control: to match the responsibilities of every key posi- 
tion with the management information necessary for the effective and 
efficient execution of those responsibilities. Control itself can be defined as 
the making of decisions and taking of actions required by the responsibili- 
ties of each position, i.e., the proper performance of each executive accord- 
ing to the requirements of his position. 

Now, some readers may object to this concept on the ground that it 
does not even mention the familiar rudiments of control, traditionally 
conceived. You may be prompted to say: "Control means making sure 
that actual results conform to desired results, and this involves three basic 
functions: (a) setting standards of satisfactory performance; (b) check- 
ing results to see how they compare with the standards; and (c) taking 
corrective action where actual results do not meet the standards." 

I have no quarrel with this concept, except that these functions ought 
to be, and in fact must be, built into the organization structure as part and 
parcel of the responsibilities and authorities of every key position. They 
should not be segregated and put on a list of functions under the heading 

This brings us to the basic law of most control systems as well as of 
most plans of organization. Control and organization have generally been 
treated independently of each other, thus missing the point of how the 
organization is to work in practice, or of what the executives are trying to 
control in the operations. Actually organization and control are insepara- 
ble when there is effective management; they cannot function properly 
without each other. 

How many times have you pored over an organization plan only to 
find yourself saying: "In general I think it is good, but how do I do this 
particular job; how can I carry out the responsibilities entrusted to me?" 
The answer here to a very large extent is in the informational and control 
system that is established. What are the "management dials" necessary for 
you to do your job? What are the significant management factors you 
should have before you in order to make executive judgment, and what is 
it you do when certain things take place? Do you know what you should 
do, and why, and what are the probable consequences of your decisions? 

This is the problem of control in every executive position in the enter- 
prise, up and down the various levels in the organization. In short, the 
answer to the effective operation of the plan of organization is primarily 
through a system of control which is part and parcel of it. 

48 Control and the Administrator 

An effective system of control, in turn, depends on the plan of organi- 
zation. If you tried to look at your control problem without considering 
the plan of organization, you would soon find yourself asking, "Who 
gives me this information? Why do I get it? What do I do with it?" Ob- 
viously you could not answer until you knew who was responsible for 
what in the scheme of things— in other words, the organization again. 

Relating control to the specific responsibilities and authorities of each 
executive position is what I have chosen to call "functional control." In 
this sense control is an integral element of every function in the organiza- 
tion, and every function will then be truly under control. 


If management accepts the concept of control just described, it will 
find that a number of challenging corollaries follow in consequence. 


First, the new concept of control will require some reorientation in the 
traditional or prevailing practices and tendencies of controllers or control 
offices. It will require recognition of the fact that a controller does not 
actually control, and that any effort on his part to take over the function 
of control from the operating personnel will lead inevitably to the abuses 
and misunderstandings I have already mentioned. 

The word "control" itself has no doubt led to the misconception that 
the control function of an enterprise is a highly centralized activity in the 
office of a controller; that management reports, statistics, and information 
generally are not only collected by the controller but are also specifically 
for his use; and, finally, that he has the authority to bring about executive 
actions throughout the organization without assuming corresponding re- 
sponsibility for those actions. 

Nothing of course could be farther from the application of effective 
control in a well-managed enterprise. Although the controller is or should 
be a major executive at the same level as the other divisional executives of 
an enterprise, and hence is or should be a part of the top-management 
team, it should be noted that, apart from his usual departmental activities 
such as keeping records, paying bills, receiving cash, preparing invoices, 
maintaining the office and routine accounting, he has no direct operating 

In fact, the more important and delicate tasks the controller must per- 
form have to do with advising the president and other executives on the 
broad, over-all picture of the enterprise; co-ordinating basic plans and 
budgets; preparing and issuing special control reports; and standardizing 

Control Means A ction 49 

methods of accounting and other procedures. Note the predominantly 
informational character of each of these responsibilities. 


The misuse of the control function has far-reaching consequences in 
the planning and budgeting process. The budget is a primary means of 
assuring that actions conform to basic plans. It is a device for measuring 
the actions taken and for determining the actions required. But it does 
not, of itself, control. 

Almost as important as the budget itself is the planning of the operation 
that is finally translated into the budget. In the planning phase, as we all 
know, actions are proposed, opposing points of view are resolved, and a 
consistent course of action is set. Conditions expected are appraised, and 
proposed actions to meet these conditions are devised. To the extent that 
planning determines the actions that need to be taken and stimulates think- 
ing ahead about those actions, it is a most effective basis for control-but, 
again, it is not control itself. 

Thus, plans and budgets together provide a picture, in common related 
terms, of what is intended and expected and the means by which the goals 
are to be achieved. They provide a means for reporting back the progress 
made against the goals, and a general framework for new decisions and 
actions in an integrated pattern of development. A good example of a 
planning and budgeting basis for control is provided by the postwar ex- 
perience of a company that was producing large, expensive precision 

The management of this company decided to prepare a report called 
a "production forecast." It was based on an estimate of sales and was 
adjusted for both engineering and manufacturing loads. When finally ap- 
proved by the executive committee it became an 1 8-month plan which 
was to be adjusted periodically in accordance with manufacturing lead 
times and in terms of what the company expected to make (the rate and 
volume of production). 

The plan was eminently successful and, even to this day, is eagerly 
awaited each quarter by all operating executives, since it has become the 
basis for many decisions throughout the company. From it can be calcu- 
lated such figures as number of direct workers required, number of work- 
ers needed in service areas (payroll and accounting), adjustments in the 
level of inventory, and adjustments in purchasing loads and selling pro- 

Prior to adoption of this plan the company was continually faced with 
unexplained increases in inventories, with serious imbalances between 
number of production workers and actual requirements, and with the 

5° Control and the Administrator 

sales department and the manufacturing department working at cross pur- 
poses. For instance, it was found that the sales department was selling 
custom designs for delivery at times when the manufacturing department 
was unable to produce even its normal load of standard lines. 

Properly conceived and used, such plans and budgets can become im- 
portant elements in implementing effective control. Rather than impeding 
judgment, they should contribute to initiative in avoiding undesirable 
conditions and in meeting such conditions when they do arise. 


Control is being exercised when the operations of the enterprise are 
guided within the plans adopted, are held in line in the face of varying 
conditions, or are returned to an in-line state after deviations are located. 
Note that action is implied in each case. This is important. In a very real 
sense, control means action— action to correct a condition found to be in 
error, or action to prevent such a condition from arising— and is never 
achieved without having action as an essential step. 

Thus, in the case of the precision equipment company whose planning 
procedure was just described, control was actually achieved through a 
series of specific steps taken by the various department heads acting in 
response to the production forecast: 

(i) Idle workers, approximating 10% of the total number, were re- 
assigned to other areas. 

(2) Personnel requirements in all service and staff areas were recalcu- 

(3) Personnel were added in the sales area in order to step up sales 

(4) Order points and ordering quantities on two major product lines 
were reset. 

(5) Deliveries on open orders with vendors were extended. 

(6) For the near future, purchasing requirements were frozen at mini- 
mum levels. 

(7) Sufficient cause was found to re-examine the company's entire in- 
ventory, which resulted in scrapping $500,000 in materials and using ma- 
terials on hand in lieu of ordering new materials. 

Note that management had first made a co-operative planning and 
budgeting effort; then came control in the form of action. 


But control action can be taken only by the individual executives who 
hold delegated responsibility and authority for the operations affected. 
Certainly it makes little sense to assign someone the "responsibility" for a 
specific operation; set the achievement of certain results as a goal; and 

Control Means Action 5 1 

then, through a series of denials, restrictions, limitations, specifications, 
and decisions, allow him no initiative. 

In such instances, the wise, loyal, and experienced executive will try to 
conform and to achieve the desired results. If he succeeds, it will probably 
be because of his own ingenuity, patience, flexibility, and doggedness, 
rather than because of any superimposed decisions coming from above. If 
he fails, however, there is a real question of accountability, since many of 
the decisions and actions will have been precipitated by others not di- 
rectly responsible for the operation. This of course is a perfect setting for 
alibis and for passing the buck when the going gets rough; someone has to 
be the scapegoat for miscalculations or poor performance. 

To sum up, merely discovering out-of-line conditions, or having de- 
tailed information about a situation, does not achieve control. Control is 
exercised by taking action, and action must be taken within the authority 
delegated. And just as no person can be said to control directly the activi- 
ties assigned to another's jurisdiction, so the only person who can directly 
control activities is the one directly responsible for them. This is funda- 
mental to the healthy and successful operation of any enterprise; at the 
same time it is probably one of the least observed principles of manage- 
ment. There are a great many more instances of its violation than there are 
of its wholehearted acceptance and practice. 


Now we can see more clearly where information, such as provided by 
the controller, fits into the picture. Many enterprises have grown beyond 
the size where they can be managed by decisions arrived at through direct 
observation alone. So there must be control; and control requires a system 
of information tailored to the specific management needs of every key 
executive— information that is timely and adequate. 

Let us take the timeliness factor first. While information as such does 
not control, it is needed by executives as a guide to actions which do con- 
trol. Because they overlook this, controllers frequently miss the mark in 
being of real service to operating personnel. In general, they submit too 
many historical reports which merely relate what has taken place. Man- 
agement would rather have approximate information that is prompt than 
highly accurate information after it is too late to be of value in decision 

Turning now to the second factor, what is the criterion of "adequate" 
information? What kind of information is adequate, and how much? Fre- 
quently controllers solve this problem by giving the line executives every- 
thing there is to know— by virtually swamping them in facts and figures. 
But the effect of too much is likely to be almost as bad as too little or 
too late. 

5 2 Control and the Administrator 

A system of control should require no more than is absolutely necessary 
in the way of reports, data, and statistics. The determination of what is 
"necessary" should conform to this simple dictum: In accord with your 
responsibilities and authority, can you or should you do anything about 
the information that is presented to you and, if so, what? This is the final 
criterion of management or control information. If you can do nothing 
about the material that is presented, then it is purely informative and not 
strictly necessary for you for management purposes. 

Every key position in an enterprise is or should be related to some ob- 
jective or set of objectives. These, in turn, should be translated into spe- 
cific goals for specific calendar periods. The executive in charge of any 
operation or department, in order to achieve these specific goals within 
certain time limits, should have before him specific performance data— 
"management dials"— so he can know how well his operation is progressing 
and, if significantly out of line, what he individually must do to correct 
the situation. 


What is needed is an understanding that the functions of planning and 
performance evaluation are part and parcel of the entire organization, and 
must therefore be distributed to each and every appropriate level of re- 
sponsibility-instead of being concentrated in a highly centralized office 
that usually carries the name of "the controller." Every key executive, in 
fulfilling his responsibilities, is or should be his own "controller." In this 
sense, control can be as much an energizing as a steering function. So con- 
ceived it should no longer be a barrier but a tremendous boon to initiative. 


Robert L. Katz* 

Mr. Robert L. Katz states that effective administration depends on three 
basic personal skills: technical, human, and conceptual. He shows the dif- 
fering degrees in which these skills are important in various levels of ad- 
ministration and discusses ways in which these skills may be further 
developed and improved. 

*From Harvard Business Review, XXXIII, 1 (1955), 33-42. Reprinted by permis- 
sion of the Harvard Business Review. 

Skills of an Effective Administrator 5 3 

Although the selection and training of good administrators is widely 
recognized as one of American industry's most pressing problems, there 
is surprisingly little agreement among executives or educators on what 
makes a good administrator. The executive development programs of 
some of the nation's leading corporations and colleges reflect a tremen- 
dous variation in objectives. 

At the root of this difference is industry's search for the traits or at- 
tributes which will objectively identify the "ideal executive" who is 
equipped to cope effectively with any problem in any organization. 

Yet this quest for the executive stereotype has become so intense that 
many companies, in concentrating on certain specific traits or qualities, 
stand in danger of losing sight of their real concern: what a man can 

It is the purpose of this article to suggest what may be a more useful ap- 
proach to the selection and development of administrators. This approach 
is based not on what good executives are (their innate traits and char- 
acteristics), but rather on what they do (the kinds of skills which they 
exhibit in carrying out their jobs effectively ) . As used here, a skill implies 
an ability which can be developed, not necessarily inborn, and which is 
manifested in performance, not merely in potential. So the principal 
criterion of skillfulness must be effective action under varying conditions. 

This approach suggests that effective administration rests on three basic 
developable skills which obviate the need for identifying specific traits 
and which may provide a useful way of looking at and understanding the 
administrative process. 


It is assumed here that an administrator is one who (a) directs the activi- 
ties of other persons and (b) undertakes the responsibility for achieving 
certain objectives through these efforts. Within this definition, successful 
administration appears to rest on three basic skills, which we will call 
technical, human, and conceptual. It would be unrealistic to assert that 
these skills are not interrelated, yet there may be real merit in examining 
each one separately, and in developing it independently. 


As used here, technical skill implies an understanding of, and proficiency 
in, a specific kind of activity, particularly one involving methods, proc- 
esses, procedures, or techniques. It is relatively easy for us to visualize 
the technical skill of the surgeon, the musician, the accountant, or the 
engineer when each is performing his own special function. Technical 

54 Control and the Administrator 

skill involves specialized knowledge, analytical ability within that spe- 
cialty, and facility in the use of the tools and techniques of the specific 

Of the three skills described in this article, technical skill is perhaps the 
most familiar because it is the most concrete, and because, in our age of 
specialization, it is the skill required of the greatest number of people. 
Most of our vocational and on-the-job training programs are largely con- 
cerned with developing this specialized technical skill. 


As used here, human skill is the executive's ability to work effectively as 
a group member and to build cooperative effort within the team he leads. 
As technical skill is primarily concerned with working with "things" 
(processes or physical objects), so human skill is primarily concerned 
with working with people. This skill is demonstrated in the way the indi- 
vidual perceives (and recognizes the perceptions of) his superiors, equals, 
and subordinates, and in the way he behaves subsequently. 

The person with highly developed human skill is aware of his own atti- 
tudes, assumptions, and beliefs about other individuals and groups; he is 
able to see the usefulness and limitations of these feelings. By accepting 
the existence of viewpoints, perceptions, and beliefs which are different 
from his own, he is skillful in understanding what others really mean by 
their words and behavior. He is equally skillful in communicating to 
others, in their own contexts, what he means by his behavior. 

Such a person works to create an atmosphere of approval and security 
in which subordinates feel free to express themselves without fear of cen- 
sure or ridicule, by encouraging them to participate in the planning and 
carrying out of those things which directly affect them. He is sufficiently 
sensitive to the needs and motivations of others in his organization so that 
he can judge the possible reactions to, and outcomes of various courses 
of action he may undertake. Having this sensitivity, he is able and willing 
to act in a way which takes these perceptions by others into account. 

Real skill in working with others must become a natural, continuous 
activity, since it involves sensitivity not only at times of decision making 
but also in the day-by-day behavior of the individual. Human skill can- 
not be a "sometime thing," Techniques canr«ot be randomly applied, nor 
can personality traits be put on or removed like an overcoat. Because 
everything which an executive says and does (or leaves unsaid or undone) 
has an effect on his associates, his true self will, in time, show through. 
Thus, to be effective, this skill must be naturally developed and uncon- 
sciously, as well as consistently, demonstrated in the individual's every 
action. It must become an integral part of his whole being. 

Skills of an Effective Administrator 5 5 

Because human skill is so vital a part of everything the administrator 
does, examples of inadequate human skill are easier to describe than are 
highly skillful performances. 


As used here, conceptual skill involves the ability to see the enterprise 
as a whole; it includes recognizing how the various functions of the or- 
ganization depend on one another, and how changes in any one part af- 
fect all the others; and it extends to visualizing the relationship of the in- 
dividual business to the industry, the community, and the political, social, 
and economic forces of the nation as a whole. Recognizing these relation- 
ships and perceiving the significant elements in any situation, the admin- 
istrator should then be able to act in a way which advances the over-all 
welfare of the total organization. 

Hence, the success of any decision depends on the conceptual skill of 
the people who make the decision and those who put it into action. When, 
for example, an important change in marketing policy is made, it is critical 
that the effects on production, control, finance, research, and the people 
involved be considered. And it remains critical right down to the last ex- 
ecutive who must implement the new policy. If each executive recognizes 
the over-all relationships and significance of the change, he is almost cer- 
tain to be more effective in administering it. Consequently the chances for 
succeeding are greatly increased. 

Not only does the effective co-ordination of the various parts of the 
business depend on the conceptual skill of the administrators involved, but 
so also does the whole future direction and tone of the organization. The 
attitudes of a top executive color the whole character of the organization's 
response and determine the "corporate personality" which distinguishes 
one company's ways of doing business from another's. These attitudes are 
a reflection of the administrator's conceptual skill (referred to by some as 
his "creative ability")— the way he perceives and responds to the direction 
in which the business should grow, company objectives and policies, and 
stockholders' and employees' interests. 

Conceptual skill, as defined above, is what Chester I. Barnard, former 
president of the New Jersey Bell Telephone Company, implies when he 
says: ". . . the essential aspect of the [executive] process is the sensing 
of the organization as a whole and the total situation relevant to it." 1 

Because a company's over-all success is dependent on its executives' 
conceptual skill in establishing and carrying out policy decisions, this skill 
is the unifying, co-ordinating ingredient of the administrative process, and 
of undeniable over-all importance. 

1 Functions of the Executive (Cambridge, Harvard University Press, 1948), p. 235. 

5^ Control and the Administrator 


We may notice that, in a very real sense, conceptual skill embodies con- 
sideration of both the technical and human aspects of the organization. 
Yet the concept of skill, as an ability to translate knowledge into action, 
should enable one to distinguish between the three skills of performing the 
technical activities (technical skill), understanding and motivating indi- 
viduals and groups (human skill), and co-ordinating and integrating all 
the activities and interests of the organization toward a common objective 
(conceptual skill). 

This separation of effective administration into three basic skills is use- 
ful primarily for purposes of analysis. In practice, these skills are so closely 
interrelated that it is difficult to determine where one ends and another 
begins. However, just because the skills are interrelated does not imply 
that we cannot get some value from looking at them separately, or by vary- 
ing their emphasis. Although all three are of importance at every level of 
administration, the technical, human, and conceptual skills of the admin- 
istrator vary in relative importance at difference levels of responsibility. 


Technical skill is responsible for many of the great advances of modern 
industry. It is indispensable to efficient operation. Yet it has greatest im- 
portance at the lower levels of administration. As the administrator moves 
further and further from the actual physical operation, this need for tech- 
nical skill becomes less important, provided he has skilled subordinates and 
can help them solve their own problems. At the top, technical skill may be 
almost nonexistent, and the executive may still be able to perform effec- 
tively if his human and conceptual skills are highly developed. For ex- 

In one large capital goods producing company, the controller was called 
on to replace the manufacturing vice president who had been stricken suddenly 
with a severe illness. The controller had no previous production experience, 
but he had been with the company for more than 20 years and knew many of 
the key production personnel intimately. By setting up an advisory staff, and 
by delegating an unusual amount of authority to his department heads, he was 
able to devote himself to co-ordination of the various functions. By so doing, 
he produced a highly efficient team. The results were lower costs, greater 
productivity, and higher morale than the production division had ever before 
experienced. Management had gambled that this man's ability to work with 
people was more important than his lack of a technical production background, 
and the gamble paid off. 

Other examples are evident all around us. We are all familiar with those 
"professional managers" who are becoming the prototypes of our modern 
executive world. These men shift with great ease, and with no apparent 

Skills of an Effective Administrator 57 

loss in effectiveness, from one industry to another. Their human and con- 
ceptual skills seem to make up for their unfamiliarity with the new job's 
technical aspects. 


Human skill, the ability to work with others, is essential to effective ad- 
ministration at every level. One recent research study has shown that 
human skill is of paramount importance at the foreman level, pointing 
out that the chief function of the foreman as an administrator is to attain 
collaboration of people in the work group. Another study reinforces this 
finding and extends it to the middle-management group, adding that the 
administrator should be primarily concerned with facilitating communica- 
tion in the organization. And still another study, concerned primarily 
with top management, underscores the need for self-awareness and sen- 
sitivity to human relationships by executives at that level. These findings 
would tend to indicate that human skill is of great importance at every 
administrative level, but notice the difference in emphasis. 

Human skill seems to be most important at lower levels, where the 
number of direct contacts between administrators and subordinates is 
greatest. As we go higher and higher in the administrative echelons, the 
number and frequency of these personal contacts decrease, and the need 
for human skill becomes proportionately, although probably not ab- 
solutely, less. At the same time, conceptual skill becomes increasingly 
more important with the need for policy decisions and broad-scale action. 
The human skill of dealing with individuals then becomes subordinate to 
the conceptual skill of integrating group interests and activities into a 
co-ordinated whole. 

In fact, a recent research study by Professor Chris Argyris of Yale 
University has given us the example of an extremely effective plant 
manager who, although possessing little human skill as defined here, was 
nonetheless very successful: 

This manager, the head of a largely autonomous division, made his super- 
visors, through the effects of his strong personality and the "pressure" he ap- 
plied, highly dependent on him for most of their "rewards, penalties, author- 
ity, perpetuation, communication, and identification." 

As a result, the supervisors spent much of their time competing with one 
another for the manager's favor. They told him only the things they thought 
he wanted to hear, and spent much time trying to find out his desires. They 
depended on him to set their objectives and to show them how to reach them. 
Because the manager was inconsistent and unpredictable in his behavior, the 
supervisors were insecure and continually engaged in interdepartmental 
squabbles which they tried to keep hidden from the manager. 

Clearly, human skill as defined here, was lacking. Yet, by the evaluation of 
his superiors and by his results in increasing efficiency and raising profits and 

5 8 Control and the Administrator 

morale, this manager was exceedingly effective. Professor Argyris suggests that 
employees in modern industrial organizations tend to have a "built-in" sense 
of dependence on superiors which capable and alert men can turn to advan- 
tage. 2 

In the context of the three-skill approach, it seems that this manager was 
able to capitalize on this dependence because he recognized the interrela- 
tionships of all the activities under his control, identified himself with the 
organization, and sublimated the individual interests of his subordinates 
to his (the organization's) interest, set his goals realistically, and showed 
his subordinates how to reach these goals. This would seem to be an ex- 
cellent example of a situation in which strong conceptual skill more than 
compensated for a lack of human skill. 


Conceptual skill, as indicated in the preceding sections, becomes increas- 
ingly critical in more responsible executive positions where its effects are 
maximized and most easily observed. In fact, recent research findings lead 
to the conclusion that at the top level of administration this conceptual 
skill becomes the most important ability of all. As Herman W. Steinkraus, 
president of Bridgeport Brass Company said: 

"One of the important lessons which I learned on this job [the presi- 
dency] is the importance of co-ordinating the various departments into an" 
effective team, and, secondly, to recognize the shifting emphasis from time 
to time of the relative importance of various departments to the business. 3 

It would appear, then, that at lower levels of administrative responsibil- 
ity, the principal need is for technical and human skills. At higher levels, 
technical skill becomes relatively less important while the need for con- 
ceptual skill increases rapidly. At the top level of an organization, concep- 
tual skill becomes the most important skill of all for successful administra- 
tion. A chief executive may lack technical or human skills and still be 
effective if he has subordinates who have strong abilities in these direc- 
tions. But if his conceptual skill is weak, the success of the whole organiza- 
tion may be jeopardized. 


This three-skill approach implies that significant benefits may result 
from redefining the objectives of executive development programs, from 
reconsidering the placement of executives in organizations, and from re- 
vising procedures for testing and selecting prospective executives. 

2 Executive Leadership (New York, Harper & Brothers, 1953); see also "Leadership 
Pattern in the Plant," HBR, January-February 1953, p. 63. 

3 "What Should a President Do?" Dun's Review, August 195 1, p. 21. 

Skills of an Effective Administrator 59 


Many executive development programs may be failing to achieve satis- 
factory results because of their inability to foster the growth of these 
administrative skills. Programs which concentrate on the mere imparting 
of information or the cultivation of a specific trait would seem to be 
largely unproductive in enhancing the administrative skills of candidates. 

A strictly informative program was described to me recently by an 
officer and director of a large corporation who had been responsible for 
the executive development activities of his company, as follows: 

What we try to do is to get our promising young men together with some 
of our senior executives in regular meetings each month. Then we give the 
young fellows a chance to ask questions to let them find out about the com- 
pany's history and how and why we've done things in the past. 

It was not surprising that neither the senior executives nor the young 
men felt this program was improving their administrative abilities. 

The futility of pursuing specific traits becomes apparent when we con- 
sider the responses of an administrator in a number of different situations. 
In coping with these varied conditions, he may appear to demonstrate one 
trait in one instance— e.g., dominance when dealing with subordinates— and 
the directly opposite trait under another set of circumstances— e.g., sub- 
missiveness when dealing with superiors. Yet in each instance he may be 
acting appropriately to achieve the best results. Which, then, can we 
identify as a desirable characteristic? Here is a further example of this 

A Pacific Coast sales manager had a reputation for decisiveness and positive 
action. Yet when he was required to name an assistant to understudy his job 
from among several well-qualified subordinates, he deliberately avoided making 
a decision. His associates were quick to observe what appeared to be obvious 

But after several months had passed, it became clear that the sales manager 
had very unobtrusively been giving the various salesmen opportunities to 
demonstrate their attitudes and feelings. As a result, he was able to identify 
strong sentiments for one man whose subsequent promotion was enthusiastically 
accepted by the entire group. 

In this instance, the sales manager's skillful performance was improperly 
interpreted as "indecisiveness." Their concern with irrelevant traits led 
his associates to overlook the adequacy of his performance. Would it not 
have been more appropriate to conclude that his human skill in working 
with others enabled him to adapt effectively to the requirements of a new 

Cases such as these would indicate that it is more useful to judge an ad- 
ministrator on the results of his performance than on his apparent traits. 
Skills are easier to identify than are traits and are less likely to be mis- 
interpreted. Furthermore, skills offer a more directly applicable frame of 

60 Control and the Administrator 

reference for executive development, since any improvement in an ad- 
ministrator's skills must necessarily result in more effective performance. 

Still another danger in many existing executive development programs 
lies in the unqualified enthusiasm with which some companies and colleges 
have embraced courses in "human relations." There would seem to be 
two inherent pitfalls here: (i) Human relations courses might only be 
imparting information or specific techniques, rather than developing the 
individual's human skill. (2) Even if individual development does take 
place, some companies, by placing all of their emphasis on human skill, 
may be completely overlooking the training requirements for top posi- 
tions. They may run the risk of producing men with highly developed 
human skill who lack the conceptual ability to be effective top-level 

It would appear important, then, that the training of a candidate for an 
administrative position be directed at the development of those skills 
which are most needed at the level of responsibility for which he is being 


This three-skill concept suggests immediate possibilities for the creating 
of management teams of individuals with complementary skills. For ex- 
ample, one medium-size midwestern distributing organization has as presi- 
dent a man of unusual conceptual ability but extremely limited human 
skill. However, he has two vice presidents with exceptional human skill. 
These three men make up an executive committee which has been out- 
standingly successful, the skills of each member making up for deficiencies 
of the others. 


In trying to predetermine a prospective candidate's abilities on a job, 
much use is being made these days of various kinds of testing devices. 
Executives are being tested for everything from "decisiveness" to "con- 
formity." These tests, as a recent article in Fortune points out, have 
achieved some highly questionable results when applied to performance 
on the job. 4 Would it not be much more productive to be concerned with 
skills of doing rather than with a number of traits which do not guarantee 

This three-skill approach makes trait testing unnecessary and substitutes 
for it procedures which examine a man's ability to cope with the actual 
problems and situations he will find on his job. These procedures, which 

4 William H. Whyte, Jr., "The Fallacies of 'Personality' Testing," Fortune, 
September 1954, p. 117. 

Skills of an Effective Administrator 6 1 

indicate what a man can do in specific situations, are the same for selection 
and for measuring development. They will be described in the section on 
developing executive skills which follows. 

This approach suggests that executives should not be chosen on the basis 
of their apparent possession of a number of behavior characteristics or 
traits, but on the basis of their possession of the requisite skills for the spe- 
cific level of responsibility involved. 


For years many people have contended that leadership ability is inherent 
in certain chosen individuals. We talk of "born leaders," "born execu- 
tives," "born salemen." It is undoubtedly true that certain people, naturally 
or innately, possess greater aptitude or ability in certain skills. But research 
in psychology and physiology would also indicate, first, that those having 
strong aptitudes and abilities can improve their skill through practice and 
training, and, secondly, that even those lacking the natural ability can 
improve their performance and effectiveness. 

The skill conception of administration suggests that we may hope to 
improve our administrative effectiveness and to develop better admin- 
istrators for the future. This skill conception implies learning by doing. 
Different people learn in different ways, but skills are developed through 
practice and through relating learning to one's own personal experience 
and background. If well done, training in these basic administrative skills 
should develop executive abilities more surely and more rapidly than 
through unorganized experience. What, then, are some of the ways in 
which this training can be conducted? 


Development of technical skill has received great attention for many 
years by industry and educational institutions alike, and much progress 
has been made. Sound grounding in the principles, structures, and proc- 
esses of the individual specialty, coupled with actual practice and experi- 
ence during which the individual is watched and helped by a superior, 
appear to be most effective. In view of the vast amount of work which 
has been done in training people in the technical skills, it would seem 
unnecessary in this article to suggest more. 


Human skill, however, has been much less understood, and only re- 
cently has systematic progress been made in developing it. Many different 
approaches to the development of human skill are being pursued by var- 

6i Control and the Administrator 

ious universities and professional men today. These are rooted Jn such 
disciplines as psychology, sociology, and anthropology. 

Some of these approaches find their application in "applied psychology," 
"human engineering," and a host of other manifestations requiring techni- 
cal specialists to help the businessman with his human problems. As a 
practical matter, however, the executive must develop his own human 
skill, rather than lean on the advice of others. To be effective, he must 
develop his own personal point of view toward human activity, so that he 
will (a) recognize the feelings and sentiments which he brings to a situa- 
tion; (b) have an attitude about his own experiences which will enable him 
to re-evaluate and learn from them; (c) develop ability in understanding 
what others by their actions and words (explicit or implicit) are trying 
to communicate to him; and (d) develop ability in successfully communi- 
cating his ideas and attitudes to others. 

This human skill can be developed by some individuals without formal- 
ized training. Others can be individually aided by their immediate su- 
periors as an integral part of the "coaching" process to be described later. 
This aid depends for effectiveness, obviously, on the extent to which the 
superior possesses the human skill. 

For larger groups, the use of case problems coupled with impromptu 
role playing can be very effective. This training can be established on a 
formal or informal basis, but it requires a skilled instructor and organized 
sequence of activities. It affords as good an approximation to reality as 
can be provided on a continuing classroom basis and offers an opportunity 
for critical reflection not often found in actual practice. An important 
part of the procedure is the self-examination of the trainee's own concepts 
and values, which may enable him to develop more useful attitudes about 
himself and about others. With the change in attitude, hopefully, there 
may also come some active skill in dealing with human problems. 

Human skill has also been tested in the classroom, within reasonable 
limits, by a series of analyses of detailed accounts of actual situations in- 
volving administrative action, together with a number of role-playing op- 
portunities in which the individual is required to carry out the details of 
the action he has proposed. In this way an individual's understanding of the 
total situation and his own personal ability to do something about it can 
be evaluated. 

On the job, there should be frequent opportunities for a superior to 
observe an individual's ability to work effectively with others. These may 
appear to be highly subjective evaluations and to depend for validity on 
the human skill of the rater. But does not every promotion, in the last 
analysis, depend on someone's subjective judgment? And should this sub- 
jectivity be berated, or should we make a greater effort to develop people 
within our organizations with the human skill to make such judgments 

Skills of an Effective Administrator 6 3 


Conceptual skill, like human skill, has riot been very widely understood. 
A number of methods have been tried to aid in developing this ability, 
with varying success. Some of the best results have always been achieved 
through the "coaching" of subordinates by superiors. This is no new idea. 
It implies that one of the key responsibilities of the executive is to help 
his subordinates to develop their administrative potentials. One way a 
superior can help "coach" his subordinate is by assigning a particular 
responsibility, and then responding with searching questions or opinions, 
rather than giving answers, whenever the subordinate seeks help. 

Obviously, this is an ideal and wholly natural procedure for administra- 
tive training, and applies to the development of technical and human 
skill, as well as to that of conceptual skill. However, its success must neces- 
sarily rest on the abilities and willingness of the superior to help the 

Another excellent way to develop conceptual skill is through trading 
jobs, that is, by moving promising young men through different functions 
of the business but at the same level of responsibility. This gives the man 
the chance literally to "be in the other fellow's shoes." 

Other possibilities include: special assignments, particularly the kind 
which involve interdepartmental problems; and management boards, such 
as the McCormick Multiple Management Plan, in which junior executives 
serve as advisers to top management on policy matters. 

For larger groups, the kind of case-problems course described above, 
only using cases involving broad management policy and interdepart- 
mental co-ordination, may be useful. Courses of this kind, often called 
"General Management" or "Business Policy," are becoming increasingly 

In the classroom, conceptual skill has also been evaluated with reason- 
able effectiveness by presenting a series of detailed descriptions of specific 
complex situations. In these the individual being tested is asked to set 
forth a course of action which responds to the underlying forces operat- 
ing in each situation and which considers the implications of this action 
on the various functions and parts of the organization and its total en- 

On the job, the alert supervisor should find frequent opportunities to 
observe the extent to which the individual is able to relate himself and 
his job to the other functions and operations of the company. 

Like human skill, conceptual skill, too, must become a natural part of 
the executive's make-up. Different methods may be indicated for develop- 
ing different people, by virtue of their backgrounds, attitudes, and ex- 
perience. But in every case that method should be chosen which will en- 
able the executive to develop his own personal skill in visualizing the 
enterprise as a whole and in co-ordinating and integrating its various parts. 

64 Control and the Administrator 


Robert M. Hutchins* 

Dr. Robert M. Hutchins comments on some problems he encountered 
in administering a university {University of Chicago) and a foundation 
(Fund for the Republic). In the process, he critically reviews a lecture he 
delivered on administration seven years before. Although the article is 
confined to "corporations not for profit," it has relevancy to administra- 
tion in general. 

I have spent 3 3 years in the administration of corporations not for profit. 
Seven years ago, at the University of Chicago, I gave a lecture on the ad- 
ministrator, and I now propose to bring that statement up-to-date and 
to offer some amendments to it. I should first like to reconsider my ex- 
perience at the University of Chicago; then I should like to make some 
observations on the relatively new field of foundation administration, and 
finally I should like to refer to the problems that have arisen to plague the 
administrator since the date of the Chicago lecture. 

That lecture suggested that what the administrator needed was the 
moral virtues and a vision of the end. I said that he had to have courage, 
fortitude, justice, and prudence or practical wisdom. I went on: "I do not 
include patience, which we are told, President Eliot came to look upon as 
the chief requirement of an administrator. ... I regard patience as a 
delusion and a snare and think that administrators have far too much of it 
rather than too little." 


I have to confess that I now believe that Mr. Eliot was right, and I was 
wrong. I now think that my lack of patience was one of my principal 
disqualifications as an administrator. 

I did not want to be an officeholder. I wanted, as the saying goes, "to 
get things done." This led me to push matters to a decision, sometimes by 
very close votes. 

It is one thing to get things done. It is another to make them last. 

I should have known that the existence of a large embittered minority, 
which felt that fundamental alterations -had been pushed through without 
consideration of its point of view, destined the alterations to endure only 
until the minority could muster the strength to become the majority. 

*From College and University Business, XIX, 5 (1955), 23-26. Reprinted by per- 
mission of the author and the F. W. Dodge Corporation. 

The Administrator Reconsidered 65 

As I said seven years ago, "The problem of time ... is insoluble. The 
administrator should never do anything he does not have to do, because 
the things he will have to do are so numerous that he cannot possibly have 
time to do them. He should never do today what he can put off till to- 
morrow. He should never do anything he can get anybody to do for him. 
He should have the largest number of good associates he can find, for they 
may be able to substitute for him. But he should be under no illusions here. 
The better his associates are, the more things they will find for him to do." 

The pressure of time is so great, the number of people who have to be 
convinced is so large, interminable discussion of the same subject with the 
same people is so boring, that the amount of patience a university ad- 
ministrator must have passes the bounds of my imagination, to say nothing 
of those of my temperament. But I have learned at last, or I think I have, 
that the university president who wants durable action, not just action, 
must have patience, and have it in this amount. 


Considerations such as these lead me to think that if I had it to do over 
again I might have begun at Chicago in 1929 with a proposal for the reor- 
ganization of the university more basic than any I ever advanced. 

The impossible size of American universities and the lamentable ex- 
tremes to which specialization has been carried lead me to believe that I 
should have proposed the reorganization of the University of Chicago on 
the lines of Oxford and Cambridge. The university should have been re- 
constituted into a federation of colleges, each representing among its stu- 
dents and teachers the major fields of learning. These colleges should have 
begun their work with the junior year, resting on the foundation of "the 
college of the university," which terminated its work at the end of the 
sophomore year. That college was intended to be the equivalent of the hu- 
manistic gymnasium or the lycee or the British public school. The 
change would have meant that basic liberal education would have been 
followed by compulsory communication with the representatives of dis- 
ciplines other than one's own throughout the whole educational process, 
and, in the case of teachers, throughout their lives. 


Such colleges as these I should have proposed— say with 250 students and 
25 faculty members— would be of manageable size. Each one could have 
an administrative officer who could be expected to lead the way to im- 
provements both numerous and lasting. The university as a whole would 

66 Control and the Administrator 

have no permanent, full-time head, The ceremonial, representative func- 
tions of the university president could be performed by a temporary of- 

The only objection I can think of to this proposal is that there would 
be no president to raise money for the university. 

The main point is that administration by persuasion and agreement, 
which is the only kind that brings lasting results, cannot be conducted in 
the vast chaos of the American university, and that the remedy for it, 
federalization, would also remedy that mutual deafness which specializa- 
tion has made the characteristic disease of the higher learning in America. 

With these important reservations and alterations, I am ready to stand 
on my lecture of seven years ago about the administration of universities. 
What about the administration of foundations, in which I have now spent 
almost five years? At first glance they seem like paradise for the university 
administrator. Foundations have no faculty, no students, no alumni, and 
no football team. The president of a foundation does not have to spend 
all his waking hours plotting how he can raise money. The Lord Chancel- 
lor of Gilbert and Sullivan said, "In my court I sit all day, giving agreeable 
girls away." Substitute millions for girls and you have the life of a founda- 
tion executive as it seems to a university president who has never been a 
foundation executive. 

When one is a foundation executive one begins to appreciate certain 
things about universities that are not always in one's ungrateful mind when 
one is administering a university. The academic tradition is so attenuated 
in this country that the administrator of an academic institution is likely 
to feel that it does not exist. His preoccupation with his trustees is likely 
to be such that he does not notice the functioning in the university of 
what John K. Galbraith has called "countervailing power." 

In a foundation, where there is no tradition yet and where there is no 
countervailing power, the administrator may recall with some nostalgia 
that the academic tradition did to some extent protect the academic enter- 
prise and that the tension among administration, faculty, trustees, students, 
alumni, donors and the public did have a tendency to keep any group 
from totally destroying the independence of the academic body. 


It is true that the academic body has been in serious danger of losing its 
independence. The reason is that nobody can understand why it should 
have it. And the reason why nobody can understand this is that the col- 
leges and universities of this country have, in their desire for popularity 
and money, gladly responded to every pressure and every demand. They 

The Administrator Reconsidered 6j 

have insisted on their dependence; they have become folk institutions, re- 
flecting the whims, no matter how frivolous or temporary, of those whose 
support they hope to gain. 

In my Chicago lecture I said, "The temptation, of course, is to bury 
yourself in routine. There is so much routine— so many reports, so many 
meetings, so many signatures, so many people to see— all of some value to 
the institution, that you can conscientiously take your salary and never 
administer at all. You can spend your entire time doing things which any 
$30 a week clerk could do better and go home at night exhausted to report 
to your wife that you have had a hard day wrestling with university 
problems. The administrator who is determined to administer will find the 
strain on his character very great." 


In foundation work the refuge of the administrator is frittering, rather 
than routine, though there is plenty of that, too. The administrator of a 
foundation has to have a vision of the end, and, in the light of it, has to 
decide what expenditure is most likely to achieve it. He must decide that 
for the purposes of his foundation A is better than B, that X University is 
better than Y University, and he must do this even though B is more in- 
fluential than A and Y more celebrated than X. The temptation here, of 
course, is to fritter, to recommend that both A and B, both X and Y, re- 
ceive some money and thus to get the best of both worlds by combining 
popularity and effectiveness. 

Most foundations, unlike the Fund for the Republic, have very general 
purposes, such as the welfare of mankind. The decision as to what ex- 
penditures will promote the real welfare of mankind is so difficult, it in- 
volves such a tremendous intellectual effort, which at best can result only 
in a guess, that a natural desire is to give nearly everybody something for 
nearly everything in the hope that some interesting entries will emerge 
for the annual report. 

But this formula will not protect you today if you give something, no 
matter how little, to somebody who is unpopular, or if you give it for an 
unpopular cause. And so I come to the third and last main division of these 
remarks. What has happened to the administration of corporations not-for- 
profit in the last few years? 

In the Chicago lecture I said, "The administrator who is afraid of any- 
body or anything is lost. The administrator who cannot stand criticism, 
including slander and libel, is lost." This is still true. I went on: "The 
natural course, then, is to become an officeholder." But even being an 
officeholder will not protect you now. The great change that has taken 

68 Control and the Administrator 

place in recent years is a change in the atmosphere in which the ad- 
ministrator has had to operate. It is an atmosphere of suspicion and fear. 

Since the clouds of suspicion and fear have been rolling in along with 
rising costs, those corporations not-for-profit which have to raise money, 
as all colleges and universities think they must, have felt constrained to 
propitiate those forces which have generated fear and suspicion or those 
persons and groups who have been influenced by it. Those corporations 
not-for-profit which do not have to raise money, like the foundations, 
have felt constrained to merge innocuously into the environment by veer- 
ing off from any activity or from any association that could be criticized. 

Now the object of universities, hospitals and foundations is not the 
preservation of the status quo. It is the improvement of the conditions of 
human life and the clarification of its aims. A university that does not 
try to improve the educational system and the environment in which it 
operates, a hospital that does not try to improve medical practice, a foun- 
dation that is not dedicated to the welfare of man is a failure. Yet uni- 
versities, hospitals and foundations that do these things must inevitably en- 
gage in criticism of existing practices, and if they do they must expect to 
be criticized in turn. 


And what about the next wind? The administrator ought to have a 
vision of the end that is clear and true regardless of meteorological 

It is not merely inevitable that we are different and have different views; 
it is desirable that this should be so. From the clash of opinion truth 
emerges, and the human race advances. 

Hence the essence of Americanism is discussion. It is not name calling or 
suppression. It is certainly not dogma or prejudice. The only political 
dogma in America is that discussion is the road to progress, that every 
man is entitled to his own opinions, and that we have to learn to live with 
those whose opinions differ from our own. After all, they may turn out to 
be right. 

The administrator must have a clear, true vision of the end, and he must 
have courage, fortitude, justice, prudence and patience in order to pursue 
it through all kinds of weather. The administrator who instead of pur- 
suing the end pursues public relations may make himself and his institution 
rich and popular. Public relations means trying to find out what the pre- 
vailing opinion is before you act and then acting in accordance with it. 

The tendency of the pursuit of public relations is to make everybody 
think, look, talk and act like everybody else: How can anybody criticize 

The Administrator Reconsidered 

6 9 

you if you are like everybody else? The pursuit of popularity brings the 
kind of success that turns to dust and ashes in the administrator's mouth; it 
means that he and his institution have failed in reality, because the end has 
got lost, and so they have not done for their fellowmen what they were 
intended to do. The moral virtues and the vision of the end— these are 
still what the administrator needs; he needs them now as never before. 



Herrymon Maurer* 

Mr. Maurer examines the development of modern management, assesses 
its present state, and speculates on its future. The phenomenon of the 
large corporation and its ramifications makes the author's article par- 
ticularly interesting and thought-provoking. 

For twenty-five years the managers of U.S. business have made it in- 
creasingly evident that this is indeed their age, not the age of enthroned 
wealth, entrenched tycoons, or enterprising manipulators of securities. 
Yet the role of managers— unlike that of the owners, magnates, and finan- 
ciers of the past— remains incompletely assayed and popularly misunder- 
stood. The older roles were settled and conspicuous. But modern manage- 
ment, self-conscious and even introspective, is engaged in constant study, 
experimentation, and change. The modern managerial breed is preoc- 
cupied, above all, with the future— not just the personal future, but the 
future of the corporation itself. 

* Reprinted from the January 1955 issue of Fortune Magazine by Special Permis- 
sion; © 1955 Time Inc. 


The Age of the Managers 7 i 

The rapid and accelerating changes of the past quarter-century have 
produced a new sort of business leader, unwilling to be propelled aim- 
lessly by the shifting winds and waves of immeasurable economic forces. 
A manager, indeed, might be defined as a man who wants to navigate his 
course of business as exactly and predictably as he can. While he makes 
various decisions by hunch, dead reckoning, or sudden reaction to cur- 
rent market conditions, he prefers to rationalize decisions, calculating the 
risks and analyzing the markets. He maneuvers not so much by feel as by 
instruments, which become more numerous— and more complex— year 
after year. The device of "linear programing" materialized only a few 
years ago, and the concept of "planning and control for profits" achieved 
wide acceptance only within the last five years. The various new tools, 
moreover, depend on other tools also relatively new: e.g., aptitude tests 
for future managers, precise cost-accounting systems, general economic 
forecasts, specific market projections. 

All such tools and all the rational calculations of managers share one 
significant characteristic: they apply more to the economic hereafter than 
to the here-and-now of business life. Indeed, modern managers are dis- 
tinguished by their concern with what profits, products, and markets will 
be in a tomorrow that may be several years to several decades distant. 
This preoccupation with futurity may prove to have good consequences 
—or bad. Will all the instruments and calculations of the modern manager 
rationalize away the responsibility, the intuitiveness, the personality, and 
even the dignity of the human beings who man large corporations? Or 
will they help the corporations to become still more capably managed, 
while remaining hospitable to the talents of individual men and women? 
The questions call for a new appraisal of the modern managers: where are 
they now, and where are they headed? 

The pacesetters of management are now, by definition, the top decision 
makers of the largest corporations. The influence of these executives is 
extensive largely because of the relationship between their corporations 
and the many smaller companies that big business tends to breed in in- 
creasing numbers. G.M., which employs over 475,000 people, is the direct 
source of employment for nearly 400,000 others in dealerships and distrib- 
utorships, and indirectly for perhaps a million people in 21,000 supply 
companies. Sears has nearly 10,000 suppliers, which employ about 
1,600,000 people. Swift's retail dealers number between 300,000 and 
400,000. Westinghouse has between 100,000 and 125,000 distributors 
and draws on the labors of an incalculable number of supplier and sub- 
contractor employees. 

Since about one out of four business employees works for the 200 
largest corporations, and since these corporations make employment for 
at least two additional persons besides every one they employ themselves, 

j 2 Decision-making and the Administrator 

it is not too much to say that the management practices of the 200 largest 
corporations strongly influence three-quarters of U.S. business life. 

1930 AND 1955 

One clue to the vast economic and social consequences of management 
activity lies in the particular quality of competition that has become prev- 
alent during the past twenty-five years. Large corporations are now in- 
creasing their size, activities, and earnings by creating new markets, in- 
creasing old ones, and developing new types of products. As regards new 
products alone, du Pont President Crawford H. Greenewalt has estimated 
that "half our present national working force is engaged in production 
and sales of things unheard of generally in 1902. A very large number are 
concerned with developments new since 1928. Should this trend continue, 
half our working population in the year 1978 may be making and selling 
things as yet unknown." 

This type of new-product and new-market competition is not of itself 
a new thing; it was always a concomitant of our expanding market. But 
more and more managers have sharpened it into a conscious policy, by 
which they set out to stimulate future consumption by planning the low- 
est possible prices on the biggest possible output. Since managers seek in 
the long run to capture the highest possible sales and profits, it is signifi- 
cant that they emphasize decreasing price as a dynamic that makes for ex- 
panding production and, as a consequence, for lower unit costs and higher 
return on investment. Companies often respond to this dynamic by elect- 
ing to take a temporary loss on certain products to increase their sales and 
consequently their production. Westinghouse, for instance, gave up the 
profit margin on its distribution transformers to build future acceptance 
for them. G.E. follows comparable practice, calls it "broadening the 

Similarly, the dynamic concept of wages, which seemed so revolu- 
tionary when Henry Ford enunciated it in 19 14, has been almost uni- 
versally accepted by the managers. Partly in response to the pressures of 
organized labor, they treat high pay not only as an incentive to employee 
productivity but as a means for turning the workingman into a consumer, 
able in most cases to buy what he makes. An expanding market coupled 
with decreasing prices and increasing production is the basis for the wide- 
spread belief among managers in an expanding economy as a natural 

Consider now the social changes that have paralleled the growth of 
these concepts of management. Since 1930 not only have the poor be- 
come richer but the rich poorer. Child labor has disappeared. The gulf 
between classes has narrowed abruptly. Big houses are smaller, small 

The Age of the Managers 7 3 

houses bigger. Servants are disappearing from the homes of the well-to- 
do, and the street and sports clothes of the assembly-line worker are 
hard to distinguish from those of executives. Employees now take for 
granted goods, products, services, standards of living, levels of education, 
and even types of architecture that not long ago would have seemed be- 
yond the reach of anyone but bosses. Meanwhile, paid vacations, medical 
benefits, insurance, and pensions have become prevalent among large 
corporations. Unionism, long fought violently, has within twenty-five 
years become a necessary part of large industrial establishments. 


More social change and marked alterations in management will prob- 
ably follow the new preoccupation of business with the future. The lag 
in time between decisions and sales is increased, of course, by the tech- 
nological complexity of new products and new plants. Nowadays, many 
managers find themselves committing funds for research, development, 
engineering design, and capital expansion that cannot produce income for 
five to ten or more years. They have to deliberate levels of price and pay, 
themselves complex considerations easily influenced by many still un- 
known factors, and they have to weigh as rationally as possible the effect 
of these levels upon distant levels of production and the even more distant 
shape of the market. Meanwhile, they have to respond as sensitively as 
possible to studies of what business conditions may conceivably be many 
months or many years hence. 

Thus managers look ahead in a more sophisticated way than they for- 
merly did. They make decisions that are considerably separated in time 
from the market transactions to which the decisions are supposed to lead. 
Du Pont spent twelve years' time and $27 million in research and develop- 
ment to get nylon into production. Before R.C.A. sold its first television 
set, it also put $50 million into research and development. Some years ago 
the Santa Fe purchased several hundred acres of land adjoining its proper- 
ties in Los Angeles after a five-year study of what the land needs of the 
railroad might be fifty years thence. It takes Consolidated Edison two 
years to plan a modern, complex generating station and from three to five 
years more to complete it. For some time Sears has been expanding its 
store selling area in the South (its square footage per capita of population 
is already greater there than in any other part of the country) out of a 
conviction that natural resources, population growth, and the stimulus of 
Sears' eleemosynary programs will make the South more prosperous fifty 
or so years hence. As early as 1929, Alcoa built a mill at Massena, New 
York, to roll large structural shapes for which there was not any market, 
and has recently built another at Davenport, Iowa, to roll unusually wide 

74 Decision-making and the Administrator 

sheets although the market cannot begin to absorb the output for some 
years to come. 

At Union Carbide & Carbon, the time lag between test tube and tank car 
is ten years for "simple" projects; but there was a gap of twelve years for 
Vinylite, more than seventeen for coal hydrogenation. At Sears, schedules 
are worked out with suppliers for products that may not be selling until 
after four, seven, or more years. Washing machine production— complete 
with program for cutting costs, improving quality, and reducing seasonal 
fluctuations— is written down in a manual that covers the next four years. 
In sum, big-company managers are engaged in a continuum of business 
decisions separated by a chasm of days, months, and years from a con- 
tinuum of market transactions. 


Solicitude for what will be, and how it will affect the long-term pro- 
duction, prices, and profits of a corporation entails, first of all, a willing- 
ness to refrain from squeezing the last drop of profit out of current 
operation. Managers often fall heir to profits resulting from decisions made 
prior to their own appointment to decision-making jobs. This circum- 
stance usually convinces them that sacrifice of maximum profits and 
dividends today is often essential to provide a higher level of profits in 
another top-executive generation. In fact, almost any manager can point 
to particular uses to which today's profits can be put for tomorrow's 
profits: research, for example. Obviously, a decision to expand plant is an 
employment of money in hand to create money in some future bush. 
Sometimes profits are plowed back and dividends held down. Sometimes 
money is borrowed and current costs increased. And always there is the 
overhead of research, engineering, development, and design. Such deci- 
sions, of course, are encouraged by the federal tax structure with its 
provisions for capital gain and rapid depreciation. 

In 1953 the plow-back rate of earnings was 50 per cent, compared with 
30 per cent in the Twenties. Even among utilities that go into the security 
market for their expansion needs, there is a readiness to shoulder the risks 
of investing great sums of money at times when costs are high. The Bell 
System, for instance, conceivably could have safeguarded its profits by 
dragging its feet when postwar demand for telephones mounted, but it 
took on a construction program that now exceeds $9 billion— more than 
the company's total assets at the end of the war. 

The managers of large corporations, moreover, usually have concrete 
plans to ensure the flow of raw materials critical to the company's future 
earnings. Witness the exploration programs of the big oil companies. Wit- 
ness also U.S. Steel's iron hunt. The corporation presumably could have 
found a more attractive return on its money than its high-cost tidewater 

The Age of the Managers 75 

Fairless Works, erected to process Venezuelan ore. But had the company 
not elected to build this plant, it would have lacked the means to compete 
with other manufacturers on the eastern seaboard. Comparable in urgency 
is American Can's intensive effort to find materials to take the place of 
tin, of which there is a free-world supply of twenty-three years if de- 
posits adjacent to Communist areas are not lost, or enough tin for just 
five years if the deposits are lost. "So far as I am concerned," remarks 
President William C. Stolk, reviewing his company's efforts to anticipate 
the future, "this is the day after tomorrow." 

Managers today do not merely look to the future, or guess at the fu- 
ture, or simply let the future come to pass. They consciously analyze, 
calculate, project, and predict. So rationalized are their decisions that they 
lead to conscious and continuous planning of the whole of a corporation's 
activities. This planning is flexible, to be sure; most programs provide for 
cutbacks at the drop of an index. But few programs can be cut out alto- 
gether. There are basic decisions, including commitments of money, that 
could be altered only by major economic reversals. When such basic de- 
cisions are being made from one to a dozen years ahead in a segment of 
U.S. enterprise potent not only in its immediate effect on the economy 
but also in its effect on an army of suppliers, subcontractors, dealers, and 
distributors, the economy no longer simply happens; it is to a very con- 
siderable extent planned. And the planning is done by the forward pro- 
grams of managers themselves. The actual planning period is stretching 
constantly: many managers, indeed, project even the ordinary events of 
their businesses and draw up detailed profit-and-loss statements for a 
rolling five-year period. 


Novel as this conscious sort of planning is, it involves other novelties, 
one of which is anticipating and creating future markets. A calculated 
market, clearly, is something different from the market places of earlier 
times, or from the commodity and security exchanges of today. Most 
corporations today are concerned not with actual places of exchange but 
with market potentials for particular products in particular areas. Thus 
there is the Midwest rural market for cake mixes or the suburban market 
for air conditioners. In competing for shares in such markets, managers 
must draw on studies of economic trends, projections of population statis- 
tics, estimates of future income within the market, and other indications 
of distant sales potentials. The remarkable fact about this uncertain sort 
of market is that most managers are confident in their anticipation of it. 

One reason for confidence is a belief by many managers that the plan- 
ning process itself may contribute to the stability and growth of the econ- 
omy and may help limit the swings of the business cycle. If G.M. has 

y6 Decision-making and the Administrator 

spent money during the past ten years that it must recover during the 
next ten; if du Pont is putting into production products that came out of 
a test tube eleven years ago; if U.S. Steel has committed itself not only to 
the costs of finding ore but also to the costs of transporting and process- 
ing it; if the Bell System has spent over $9 billion on expansion largely 
to anticipate future demand; if, in short, several hundred large companies, 
together with the small companies grouped around them, have gone to 
great expense in planning future products and plants and have taken great 
pains to calculate the chances of this expense producing future profits, 
then a significant stabilizing force has been brought to bear on the 

Since flexibility is a purposely built-in feature of most plans, managers 
can tailor part of their programs to the emerging shape of the economy. 
There remain, however, incalculable economic forces. It may be, for 
instance, that the heavy use of profits for capital expansion in recent years 
reflects stockholder interest in capital gains rather than in dividends; and 
it is conceivable that more expansion has occurred than the markets for 
goods justify. On the other hand, the caution with which managers tend 
to calculate future markets probably reduces the likelihood of such a 
dangerous imbalance. 

Ask a manager how he actually makes plans and he will describe how 
his company is organized. While rational organization of the labor of 
many employees is a long-established fact of large enterprises, the organi- 
zation of executives is recent. In most companies, indeed, the shift from 
one-man rule, or undefined rule, has occurred within the last twenty-five 
years. Authority and responsibility have been decentralized to the end that 
executives down the line may know what role they play in current activi- 
ties and in future planning. At the same time, top managers are putting 
new emphasis on the function of central control, trying to find the correct 
balance between decentralization and integration of management. 


Modern organization systems are seldom final. Just as managers are 
forever planning the future, so are they forever reorganizing themselves. 
But there is one crucial principle that does not change: on every executive 
level, many of the decisions are the result of the interaction of many men, 
even if the decisions are formalized by individual managers. Group man- 
agement has become characteristic of the large corporation. The corpora- 
tion's activities, particularly its future plans, are simply too vast for one 
man to keep in his head. Union Carbide, for instance, acts on the premise 
that major decisions for the future demand collective judgment. President 

The Age of the Managers yj 

Dial presides at committee meetings, but decisions result from an inter- 
action of minds, with consequent compromises, and major commitments 
are always unanimous. "When you have more new things than you have 
money," he points out, "you don't do anything unless the reasons are 
overpowering and you all agree." 

It follows that the job of such a manager as Morse Dial is to use the 
ideas of his fellow managers. In that role there is only limited scope for a 
man to practice some previous specialty. In decades to come, the U.S. is 
likely to see more and more managers who are not so much experts in law 
or finance or production or research, as in the art of management itself, 
and who still know how to get the highest volumes of sales and profits out 
of highly specialized undertakings. Of B. F. Goodrich's President William 
Richardson, a manager without formal training in chemistry, Sidney 
Weinberg has said, "I don't know any man who knows more about the 
chemical business." And du Pont's Greenewalt has noted that "specific 
skill in any given field becomes less and less important as the executive ad- 
vances through successive levels of responsibility. . . ." 


Compared with businessmen twenty-five years ago, today's managers 
exercise less individual power, definitely take home less money, probably 
enjoy less popular recognition. Generally speaking more money is to be 
had by entertainers, more power by men in government, more recognition 
by athletes. Under such circumstances, what motives make managers 
work— work harder, longer, and with greater dedication than the men of 
twenty-five years ago? The motives are unquestionably mixed, ranging 
from money and success to satisfaction in making mammoth organizations 
tick. But many top managers profess an additional motive growing out of 
the group deliberation that so often decides future business plans. For 
good or ill, the corporation is frequently the executive's club, and to some 
extent (or so his family complains) his home. Leroy Wilson, late president 
of A. T. & T., liked to stress the satisfactions of working with like-minded 
people. Union Carbide executives actually refer to their company as "the 
community." James T. Leftwich, president of F. W. Woolworth, says 
simply, "We live Woolworth." 

This bond to the company is reinforced by a significant trend toward 
setting recompense for managers essentially by policy. There is a declin- 
ing turnover of top executives in most industries (notable exceptions: air- 
craft, textile, amusements), and recompense can seldom be set in terms of 
competitive bidding for talent. It is set most often in terms of "what the 
job is worth." This development reflects the increasing confinement of 

«g Decision-making and the Administrator 

executives to the preserves of their own companies. And the confined ex- 
ecutive may well be the sort of man who centers his entire life around his 
company, from which alone he is certain of employment, and to which 
alone he is able to look for old-age income. The policy-set salary thus be- 
comes a symbol of a community closed to outsiders. 

In many ways the development of a community of executives can be 
salutary. The corporation ceases to be a mental abstraction or— what it is 
legally supposed to be— a fictional person, devoid of spirit and life. It be- 
comes a group that gives a man the sense of being an integral and valuable 
part of a common effort. The group, moreover, meets his money needs, 
offers legal advice, medical service, loans when needed, advanced educa- 
tion, even travel. It takes note of the birth of children, regularly attends 

Group management, however, has pitfalls. Big companies already have 
unique traits and habits, so that it is possible to speak of a G.E. man or a 
G.M. man or a Bell System man (once labeled homo telephonicus) . Co- 
hesiveness is one of the certainties of life at B. F. Goodrich, top money- 
maker among rubber companies per dollar of sales. "We all work so 
closely," says one Goodrich man, "that we know what each other is think- 
ing." If such closeness of purpose should lead to dull uniformity of mind, 
there would be danger: obvious danger to the profits and productivity of 
American industry because of closed-mind business thinking wherein each 
executive would repeat a line of ideas common to the group, and wherein 
no executive would offer fresh slants or new flashes of insight. A more in- 
sidious danger would be the violation of the democratic belief that the 
individual human being is unique, unrepeatable, and no man's copy. 

These dangers, to be sure, do not show up conspicuously among con- 
temporary managers, most of whom have experienced the rough-and- 
tumble of individualistic enterprises in times past. Young executives, how- 
ever, obviously lack this experience. There are signs, moreover, that some 
companies select and advance young executives partly on their ability 
(and even that of their wives) to conform to a company's pattern. There 
are also signs that an increasing proportion of younger managers, unlike 
their elders, are sons of men who were themselves executives. Thus there 
arises the unhealthy possibility of a self-perpetuating managerial elite. 


Group enterprise, indeed, offers alternatives both hopeful and horrible. 
It is conceivable that such enterprise will be the prototype of social 
slavery, that it will be the agent of a soul-rotting decay that shreds the 
last remnants of man's dignity. But it is also conceivable that the large 
corporation will emerge as a new social force whose basic drive is the 

The Age of the Managers 79 

creativity of individual human beings. The American people will, of 
course, make the choice between these alternatives, but the actual shaping 
of the alternatives will in large part be the work of tomorrow's managers. 
Some of the managerial decisions will have to do with the forms of 
group enterprise itself. Others will center on more diverse social responsi- 
bilities. Already the new management provides or extends employee 
benefits, supports charitable and educational activities, works at cut-rate 
profits for the government-a dollar a year in the case of du Pont's and 
G.E.'s atomic projects. There have been experiments with guaranteed 
annual pay, with wage increases based on productivity. There are an 
increasing number of programs to bring employees as well as executives 
into the corporate community. Such a program at Jersey Standard has 
helped maintain peaceable labor relations for more than thirty years. 


The question for the future is how well the new management, and the 
American people, understand the relationship between these social re- 
sponsibilities and the economic functions of the modern corporation. And 
here the beginning of wisdom lies in two propositions: 

All the "strictly economic" functions of management have weighty 
social consequences. 

The ability of modern management to carry out its "social" responsibili- 
ties rests on economic performance— i.e., profitability. 

It is futile for a big corporation to talk of being "a good citizen" if it is 
not meeting the first test of corporate citizenship: production and distri- 
bution at such prices and in such volume as to elevate the U.S. standard 
of living-on profit margins that assure its ability to attract capital, raise 
productivity, and yet again enhance the standard of living. It is equally 
unrealistic for a big corporation to pretend that the market is the only 
payoff on its activities; a corporate decision to have no social policies is 
actually an economic decision with adverse effects on profits. 

Much of U.S. management would subscribe to the foregoing, but few 
managers are entirely free of embarrassment in explaining to employees 
or the public the "social" value of profits or in explaining to fellow man- 
agers the "economic" value of decency. This slowness in recognizing the 
essential unity of a company's multitudinous affairs suggests that managers 
themselves are not yet fully aware of the logic of the large corporation 
with its emphasis on doing now whatever will contribute most to the 
health (i.e., profits) of the corporate community in the future. 

This difficulty is important because a substantial part of the American 
people adhere to stereotypes about managerial enterprise even though the 
stereotypes are only rough caricatures of business activity twenty-five or 

8o Decision-making and the Administrator 

more years ago. Many people still suppose that managers are the rapacious 
spokesmen of "vested interests" threatening the economic stability of the 


Obviously, the persistence of such misunderstanding means that large 
corporations, only recently rehabilitated in public esteem, are still po- 
tentially subject to attack in the future, conceivably to increased govern- 
ment control. Today's managers cannot be held wholly responsible for the 
misunderstanding. Managerial enterprise is too new to be understood 
quickly and fully, let alone explained. The responsibility therefore falls 
upon tomorrow's managers. Certainly it should not take many more years' 
experience in the logic of the large corporation to find ways to explain 
the dynamism and stability that managerial planning brings to the econ- 
omy, or to make clear that the strenuous competition vital to productive 
efficiency is not of a dog-eat-dog character. 

Nor should it be difficult to demonstrate that large corporations are 
subject not only to law but also to effective economic controls: first, com- 
petition itself, which ensures high production at low prices and also 
measures efficiency and foresight; second, the economic vote of custom- 
ers, to which managers must respond sensitively if they want to stay in 
business; third, the major test of competence, profits on investment; fourth, 
public opinion. 

Nor should tomorrow's managers find it hard to demonstrate the con- 
cern of big companies for social values. The large corporation today is not 
out of harmony with a papal encyclical (considered radical when issued 
in 1 891) which declared that a market wage was not necessarily a just 
wage, voiced approval of unions, and emphasized that people other than 
property owners had needs. Nor is the corporation out of harmony with 
comparable but unofficial statements published in the late forties by the 
Federal Council of Churches. Managers would agree that man is not a 
mere economic creature, that labor is not a mere commodity, that the 
fullest possible employment is a social good, that prices should stimulate 
consumption, and that workers should get more than subsistence pay. 


How well tomorrow's managers will do their social and economic jobs 
depends on the sort of men— alert, creative individuals or conforming 
robots— that managers now in office are choosing to succeed them. And 

Managerial Decision-making 8 1 

making the right choice is clearly complicated by the existence of group- 
minded executive communities, many of which— fortunately not all— are 
in effect closed to outsiders. 

Here is an immediate opportunity that can have far reaching benefits. 
Today's managers can reverse the long trend toward advancement strictly 
from the inside, outlaw the executive closed shop, and re-establish an 
executive market in which their various companies can actually bid for 
talent. Today, when the art of management is more important than special 
skills, executives ought to be free to circulate more easily than in the past. 
And such circulation as does occur indicates that neither the group de- 
liberation not the esprit typical of the new management would suffer from 
occasional infusions of fresh talent. Indeed, the first job of the present 
generation of managers may well be to keep the future managers out of 
institutional ruts. 


Robert TannenbaunV 

Professor Tannenbaum explores the fundamental nature of decision- 
making and the factors influencing its practice, behavior, and acceptance. 
An article by Professor Tannenbaum (on p. 16) may be regarded as a 
■first part or introduction to the present one. 


Human behavior results from either unconscious or conscious processes. 
When these processes are conscious, decision-making is involved. Deci- 
sions, when made, affect the behavior of an individual. An individual may 
make decisions which affect his own behavior, or he may make them to 
affect the behavior of another or others. In the latter case social processes 
are involved. 

Managers are those who use formal authority to organize, direct, or 
control responsible subordinates in order that all service contributions be 
co-ordinated in the attainment of an enterprise purpose. One of the most 

♦From The Journal of Business, XXIII, i (1950), 22-39. Reprinted by permission 
of the University of Chicago Press. 

82 Decision-making and the Administrator 

important techniques of managers is that of decision-making. This tech- 
nique pervades the performance of all the functions of managers. 1 In 
order to organize, direct, or control responsible subordinates, managers 
must make decisions which affect the behavior of those subordinates. The 
decisions of managers are made not to affect their own behavior but 
rather that of others. On the other hand, nonmanagers, in performing their 
work, must also make decisions, but these decisions affect only their own 

1 Robert Aaron Gordon has developed the following incomplete, but informative, 
classification of the more important types of business decisions: 

"I. Promotion and initial organization 
"A. Determination of main objectives 
"B. Setting up initial organization, involving: 
"i. Decisions as to size, legal organization, financial structure, internal organiza- 
tion, specific products to be produced and methods of producing and dis- 
tributing them, and so on 
"2. Choice of key personnel 
"C. Negotiation for the hire of the factors of production, particularly capital 
"II. Existence as a going concern 

"A. Maintenance of organization through personal leadership and continuous 

exercise of authority 
"B. Determination of the more important decisions relating to: 
"1. Volume of output and control of production 
"2. Prices '^ 

"3. Marketing (sales organization and methods, advertising, purchasing, and so 

"4. Wages and other labor problems 
"5. Financial problems, such as changes in capital structure, maintenance of 

working capital, securing new funds, and so on 
"6. Changes in the size of the firm (expansion or contraction) 
"7. Changes in the location of the firm or of important branches 
"8. Changes in internal organization and procedures 
"9. Changes in products and in (technical) methods of production 
"10. Relations of the firm with outside groups, either with specific groups, 
such as consumers, bankers, or government, or with the public in general 
(that is 'public relations') 
"11. Distribution of profits 
"C. Choice of the men who will make the above decisions and also of those pri- 
marily responsible for directing the execution of them 
"III. Reorganization or liquidation 
"A. Reorganization 
"1. Decision to reorganize 

"2. Determination, for the transition from the old to the reorganized firm, of 
the matters listed under I 
"B. Liquidation 

"1. Decision to liquidate 

"2. Deciding the terms for the liquidation, primarily with respect to the 
distribution of assets" 
(Business Leadership in the Large Corporation [Washington: Brookings Institution, 
1945], pp. 53-55.) 

And Edwin O. Stene has said: "Decision is necessary whenever an organization is 
formed, whenever routine interactivity is deliberately changed, and whenever action 
is called for which has not become routine in character" ("An Approach to a Sci- 
ence of Administration," American Political Science Review, XXXIV [December, 
1940], 1130). 

Managerial Decision-making 83 

behavior. The decisions of managers have a social import; those of non- 
managers, an individual import. Furthermore, decisions are made by other 
individuals and groups which affect the behavior of managers. These de- 
cisions also have a social import. 

In the discussion which follows, primary attention is given to decision- 
making in its social context, since this is the context which is relevant in 
so far as an understanding of the work of managers is concerned. First, 
the nature of decision-making will be analyzed. Then the interindividual 
and intergroup relationships which make it possible for the decisions of 
one to affect the behavior of another will be explored. And, finally, the 
conclusions will be related to the work of managers, indicating how man- 
agers affect the behavior of their subordinates and how others affect the 
behavior of managers. 


Etymologically, "to decide" means "to cut off." In its present usage it 
suggests the coming to a conclusion. It "presupposes previous considera- 
tion of a matter causing doubt, wavering debate, or controversy and im- 
plies the arriving at a more or less logical conclusion that brings doubt, 
debate, etc., to an end." 2 

Decision-making involves a conscious choice or selection of one be- 
havior alternative from among a group of two or more behavior alterna- 
tives. 3 In making a decision, an individual must become aware of relevant 
behavior alternatives, define them, and finally evaluate them as a basis for 
choice. To understand clearly what is involved in the making of a deci- 
sion, it will be helpful carefully to examine each of these steps in the 
decision-making process. 

/. Awareness of behavior alternatives .—Before making a decision, an 
individual should become aware of all those behavior alternatives which 

2 "Decide," Webster's Dictionary of Synonyms. 

3 Various attempts have been made to define "decision." Examples of these are: 
"The selective determination of an end for action by choice between alternatives" 
("Decision," Dictionary of Philosophy and Psychology, ed. James M. Baldwin, Vol. 
I [1940]); "Decision is the conscious consideration and conclusion regarding a course 
of action" (Stene, op. cit., p. 11 30); "A preliminary to conscious activity is a decision 
between alternatives— to do this or to do that, to do or not to do. In the process of 
decision-making the individual assesses a situation in the light of these alternatives. A 
choice between values congenial to the larger value-system of the individual is some- 
how reached" (R. M. Maclver, Social Causation [Boston: Ginn & Co., 1942], p. 296); 
"At any moment there are a multitude of alternative (physically) possible actions, 
any one of which a given individual may undertake; by some process these numerous 
alternatives are narrowed down to that one which is in fact acted out. The words 
'choice' and 'decision' will be used interchangeably ... to refer to this process. Since 
these terms as ordinarily used carry connotations of self-conscious, deliberate, ra- 
tional selection, it should be emphasized that as used here they include any process 
of selection, regardless of whether the above elements are present to any degree" 
Herbert A. Simon, Administrative Behavior (New York: The Macmillan Co., 1947). 

84 Decision-making and the Administrator 

are relevant to the decision to be made. But this is seldom, if ever, possible. 
To a considerable extent he must depend upon his own limited experience 
and information. And memory of these is often sketchy and incomplete. 
He can discover relevant behavior alternatives through investigation, by 
tapping the experience and knowledge of others. But this process is often 
excessively time-consuming and does not guarantee complete coverage of 
all alternatives. For these reasons, it is exceedingly doubtful whether most 
decisions are based upon an awareness of all relevant behavior alternatives. 

2. Definition of behavior alternatives.— -Once the individual has become 
aware of certain behavior alternatives, he is next faced with the problem 
of defining each of them. Ideally, this definition involves a determination 
of all the consequences related to each behavior alternative under consid- 
eration; but this ideal can never be achieved for the following reasons: 
(a) The most significant characteristic of the behavior alternatives is that 
their consequences lie in the future and therefore must be anticipated. But 
whenever the future is anticipated, uncertainty is present. This uncer- 
tainty is present for two reasons. In the first place, an individual never has 
the knowledge to make it possible for him accurately to determine the 
nature of the consequences which will follow upon the choice of a given 
behavior alternative or their probability of occurring, assuming that all 
other related elements remain constant. And because he does not have 
knowledge of the future, he must use imagination in attaching values to 
the consequences, which values may not obtain when the consequences 
are actually experienced. In the second place, all other related elements 
will not remain constant, (b) It is impossible for an individual to be aware 
of all the consequences attendant upon any given behavior alternative. 
(c) The time involved in discovering consequences and determining their 
nature is often such that a decision must be made before all the foreseeable 
relevant possibilities can be explored. 

5. Evaluation of behavior alternatives .—After an individual has become 
aware of certain behavior alternatives and has considered many of the 
consequences attendant upon each of these alternatives, he must next 
exercise a choice between them, i.e., make a decision. What can be said 
of the mental processes which culminate in decision? 

The decisions which an individual makes are basically of two types. 
Some (a very small proportion) of his decisions relate to his system of 
values— they determine his ultimate ends. All other decisions are directly 
or indirectly related to means for the attainment of these ultimate ends. 
The adjective "ultimate" is used advisedly. An individual's behavior is 
guided by innumerable intermediate ends, for each one of which there are 
related means. And the end of one means-end nexus becomes a means to 
a higher-order end. Decisions relating to ultimate ends cannot be judged 

Managerial Decision-making 85 

as to their efficacy. They have primarily an ethical content. But such is 
not the case with all other decisions. These are made in terms of related 
intermediate ends. In choosing between alternatives, a rational individual 
will attempt to make a selection, within the limits of his knowledge, which 
will maximize results (the degree of attainment of the relevant end) at a 
given cost or which will attain given results at the lowest cost. Thus, he 
has a criterion to guide his choice-the criterion of rationality. 

There are definite limits, however, to rational behavior viewed objec- 
tively (i.e., from an omniscient point of view). These limits stem from 
the individual's lack of knowledge with respect to the existence of be- 
havior alternatives and the consequences that will follow from them 
both from the subjective processes which are necessarily involved in de- 
fining alternatives when uncertainty is present, from time limitations, and 
from the psychological difficulties involved in holding alternatives and 
their consequences in focus preparatory to making a decision. Because 
of these factors, it is next to impossible to describe the mental processes 
which culminate in decision. 4 

The necessity for making decisions arises out of the fact that knowledge 
of relevant existing facts is inadequate and that the future is uncertain- 
individuals can never have complete knowledge of all factors underlying 
their choices. If such knowledge were available, decisions would not have 
to be made. If an individual were aware of all the consequences related 
to each of these behavior alternatives, judgment would not have to be ex- 
ercised. One alternative would clearly be superior to all others. Individual 
behavior could be completely rational. In a real sense, that behavior would 
be determined by the consequences related to the superior alternative 
rather than by a choice between alternatives. The relationship of un- 
certainty to decision-making has been stated by Frank H. Knight as 

4 On this matter, Frank H. Knight has made the following observations: "The 
mental operations by which ordinary practical decisions are made are very obscure, 
and it is a matter for surprise that neither logicians nor psychologists have shown 
much interest in them. Perhaps (the writer is inclined to this view) it is because there 
is really very little to say about the subject . . . when we try to decide what to ex- 
pect in a certain situation, and how to behave ourselves accordingly, we are likely to 
do a lot of irrelevant mental rambling, and the first thing we know we find that we 
have made up our minds, that our course of action is settled. There seems to be very 
little meaning in what has gone on in our minds, and certainly little kinship with the 
formal processes of logic which the scientist uses in an investigation. We contrast 
the two processes by recognizing that the former is not reasoned knowledge, but 
'judgment,' 'common sense,' or 'intuition.'" (Risk, Uncertainty, and Profit ["Series 
of Reprints of Scarce Tracts in Economic and Political Science," No. 16 (London: 
London School of Economics and Political Science, 1933)!. PP- 267 f.). Likewise, 
Maclver states: "There are unfathomed psycho-organic processes involved -in the 
formulation of the dynamic alternatives, in the choice between them, and in the 
passage of decision into action" (op. cit., p. 333). 

86 Decision-making and the Administrator 

. . . With uncertainty absent, man's energies are devoted altogether to doing 
things; it is doubtful whether intelligence itself would exist in such a situation; 
in a world so built that perfect knowledge was theoretically possible, it seems 
likely that all organic readjustments would become mechanical, all organisms 
automata. With uncertainty present, doing things, the actual execution of 
activity, becomes in a real sense a secondary part of life; the primary problem 
or function is deciding what to do and how to do it. . . . 5 

One further point demands attention in connection with this discussion 
of the nature of decision-making. What initiates the decision-making 
process? At any given moment of time, there are often many problems 
which might compete for an individual's attention. What determines the 
particular problem with which he will deal? Simon points out that deci- 
sion-making is initiated by stimuli, internal or external to the individual, 
which channel his attention in definite directions. Very often these stimuli, 
impinging upon the individual, are accidental and arbitrary in character. 
To the extent that they are, the individual's behavior cannot be rational. 
Also, since the attention-directing stimuli can be external to the individual, 
they can be provided by others who desire to affect the individual's 

The preceding discussion of the nature of decision-making has, for the 
most part, dealt with decision-making in the abstract. It has indicated the 
limits to rational behavior on the part of the relatively isolated individual 
—limits which are greatly reduced when the individual is a member of a 
group, as will be pointed out later. Decision-making actually takes place 
in an environment which significantly affects the decision-making process. 
In the two sections which follow, the various aspects of this environment 
will be explored in considerable detail. 


At the beginning of the preceding section it was stated that an individual 
may make decisions which affect his own behavior, or he may make them 
to affect the behavior of another or others. The latter case, as was pointed 
out, is the relevant one in so far as the work of managers is concerned. 
This relevancy is present for two reasons. First, in order to organize, di- 
rect, and control responsible subordinates, managers must make decisions 
which affect the behavior of those subordinates. And, second, managers 
themselves are in a subordinate position both with respect to their own 
superiors and formal subordinates within an enterprise and often with 
respect to others. Thus managers make decisions affecting the behavior 
of subordinates at the same time that decisions are being made which af- 
fect their own behavior. The concern here, then, is with social processes. 

5 Op. cit., p. 268. 

Managerial Decision-making 87 

These processes are those of authority and influence. In this section the 
discussion will be devoted to an analysis of these two processes. 


A superior is able directly to affect the behavior of a subordinate if he 
possesses authority with respect to that subordinate. In the preceding 
article it was stated that authority is commonly viewed as originating at 
the top of an organizational hierarchy and flowing downward therein 
through the process of delegation. When viewed in this way, it was called 
"formal authority." In reality, effective authority does not originate in 
this manner. 

The real source of the authority possessed by an individual lies in the 
acceptance of its exercise by those who are subject to it. It is the sub- 
ordinates of an individual who determine the authority which he may 
wield. Formal authority is, in effect, nominal authority. It becomes real 
only when it is accepted. An individual may possess formal authority, but 
such possession is meaningless unless that authority can be effectively 
used. And it can be so used only if it is accepted by that individual's sub- 
ordinates. Thus, to be effective, formal authority must coincide with au- 
thority determined by its acceptance. The latter defines the useful limits 
of the former. 

The concept "authority," then, describes an interpersonal relationship 
in which one individual, the subordinate, accepts a decision made by 
another individual, the superior, permitting that decision directly to affect 
his behavior. An individual always has an opportunity, with respect to a 
decision made by another directly to affect his behavior, to accept or re- 
ject that decision. If he accepts it, he thereby grants authority to its for- 
mulator and, for this matter, places himself in the position of a subordi- 
nate. As a subordinate, the individual permits his behavior directly to be 
affected by the decisions of his superior. If the individual rejects the de- 
cision, he does not grant authority to its formulator. Thus the sphere of 
authority possessed by a superior is defined for him by the sphere of ac- 
ceptance of his subordinates. 

If this line of analysis is to be followed it must be recognized that an 
individual may possess authority in a given situation without having 
formal authority. In other words, the channels through which effective 
authority is exercised do not have to follow the lines of formal organiza- 
tion within a given complex. And these channels may extend outside the 
given complex. 


Since the sphere of authority possessed by a superior is defined for him 
by the sphere of acceptance of his subordinates, it is important to inquire 
into the factors which determine this latter sphere. Why do subordinates 

88 Decision-making and the Administrator 

accept, rather than reject, the authority of their superiors? In answering 
this question, it must be remembered that the choice between acceptance 
or rejection involves a decision between two alternatives. This choice is 
made only after the individual has appraised, to the extent possible, the 
consequences attendant upon each of these alternatives. 

An individual will accept an exercise of authority if the advantages ac- 
cruing to him from accepting plus the disadvantages accruing to him from 
not accepting exceed the advantages accruing to him from not accepting 
plus the disadvantages accruing to him from accepting; and, conversely, 
he will not accept an exercise of authority if the latter factors exceed the 
former. Thus a decision to accept or reject a given exercise of authority 
results from a relative evaluation of the consequences— both positive and 
negative— attendant upon the choice of each of the competing behavior 
alternatives. To understand better the factors underlying a decision to 
accept or reject, it will be helpful to consider in more detail the nature of 
the positive and negative consequences— the advantages and disadvantages 
—related to each behavior alternative. 

The possible advantages accruing to an individual from accepting a 
given exercise of authority are many. While the following listing of types 
of advantages is by no means complete, it will serve the end of indicating 
the variety of such advantages: (i) By accepting an exercise of authority, 
an individual is able thereby to contribute to the attainment of an enter- 
prise purpose which he recognizes as being good. 

It was pointed out that group activity involves the specialized service- 
contributions of individuals which must be co-ordinated in the attainment 
of an enterprise purpose. Such co-ordination can be achieved only through 
the exercise of authority on the part of the individual who heads the 
group. An awareness of this necessity leads individuals who recognize 
the enterprise purpose as being good to accept authority. (2) By accepting 
an exercise of authority, an individual may thereby attain the approbation 
of his fellow-workers. For most individuals, social acceptance is a strong 
motivating factor. (3) By accepting an exercise of authority, an individual 
may thereby obtain rewards from his superior. These rewards might be 
increased pay, promotion, prestige, opportunity for increased personal 
power, and the like. (4) By accepting an exercise of authority, an indi- 
vidual may thereby be acting in accordance with his own moral standards. 
Some individuals believe they ought (that it is right) for them to obey 
duly constituted authorities. (5) By accepting an exercise of authority, an 
individual may thereby avoid the necessity of accepting responsibility. 
(6) Finally, by accepting an exercise of authority, an individual may 
thereby be responding to the qualities of leadership exhibited by his 
superior. In part, this point overlaps some of the preceding ones, but it 
also includes a recognition of the fact that some individuals obey others 

Managerial Decision-making 89 

out of respect for their age, superior ability or experience, character, 
reputation, personality, and the like. 

An individual, after considering the advantages to him, may often de- 
cide to accept an exercise of authority. Such a choice would be a free 
one— one not involving compulsion. But he might also decide to accept an 
exercise of authority even though the advantages attendant thereto, taken 
alone, would not be sufficient to induce him to accept. In this case his 
decision would be compelled by another or others; he would simply ac- 
quiesce to authority because of the disadvantages accruing to him from 
not accepting. When an individual is forced by another to do something 
against his will— something he otherwise would not have done— coercion 
is involved. Horace M. Kallen has defined and elaborated upon the con- 
cept of coercion in these terms: 

Coercion as a trait of human behavior may be said to obtain wherever action 
or thought by one individual or group is compelled or restrained by another. 
To coerce is to exercise some form of physical or moral compulsion. . . . 

. . . When it [coercion] is direct we call it physical; when indirect, moral. 
Both compel or restrain conduct by force majeure. . . . 

Most social coercion is indirect; it only threatens force. . . . But all coer- 
cions involve fear of penalties. Without belief that the coercer can and will 
impose penalties no indirect coercion can be effective. . . . 6 

There are numerous coercive devices the actual use or fear of which is 
often effective in obtaining an acquiescence to authority. Some examples 
of these are social disapprobation, expulsion from a group (ostracism), 
formal disciplinary action, exertion of economic pressure (monopolistic 
and monopsonistic power), torture, imprisonment, taking of a life, etc. 

In order to understand the full implications of coercion, it will be useful 
briefly to digress in order to contrast this concept with the concept of 
sanctions. In every social group there are certain modes of behavior which 
are generally approved and others which are generally disapproved. The 
reactions of approval and of disapproval represent general formal or in- 
formal social consensus. Now in every group there are some individuals 
who have urges toward nonconformity. Sanctions are devices used to in- 
duce these individuals to conform to the group will. Sanctions may be 
positive or negative in character. Positive sanctions are the social reward 
of conformity, and negative sanctions are the social consequences of non- 
conformity. All negative sanctions are coercive in their effect— they are 
used to impel an individual to conform against his will to group norms 
of behavior. Heads of groups who impose negative sanctions on recal- 
citrant individuals have the support of the vast majority of the group in 
so doing. But heads of groups might also induce conformance to behavior 
patterns which are not generally approved by the group. Here coercion 

6 "Coercion," Encyclopaedia of the Social Sciences, Vol. Ill (1930). 

go Decision-making and the Administrator 

is involved but not negative sanctions. Furthermore, one individual with 
respect to another individual or a group or one group with respect to 
another group or an individual outside the group might induce con- 
formance to specified behavior patterns through the use of coercion. But 
here again negative sanctions are not involved. Thus coercion includes 
more than negative sanctions. The latter term is applicable only when 
general group consensus is involved. When the head of a group uses coer- 
cion not based upon general group consensus, he is acting in an autocratic 
or arbitrary manner. 

In raising the question as to whether an individual will accept an exer- 
cise of authority, only the advantages accruing to him from accepting 
plus the disadvantages to him from not accepting have thus far been con- 
sidered. But he will accept only if these factors exceed the advantages 
accruing to him from not accepting plus the disadvantages accruing to 
him from accepting, as has previously been pointed out. These latter 
factors, however, are the same in nature as the former ones. An illustration 
should suffice to make this clear. An employee may be a member of a 
group most of whose members are restricting output in opposition to a 
wage-incentive plan. The implied or explicit order of the employee's 
boss is that each employee should produce the maximum output reason- 
ably possible. Should the employee accept this exercise of authority, or 
should he restrict his output (i.e„ accept the authority of the work 
group)? If he accepts the authority of his boss, he can earn more, he may 
get a desired promotion, etc. If he does not accept, he may be demoted or 
fired. On the other hand, if he does not accept, he may receive the ap- 
probation of his fellow-workers, additional status in the group, etc. And 
if he accepts, he may receive the disapprobation of his fellow-workers 
or be ostracized from the group. His decision to accept or not accept the 
boss's exercise of authority will be based upon an evaluation of these and 
similar relevant consequences. 

In order to make complete this discussion of the determinants of the ac- 
ceptance of authority, one further point calls for attention. Authority is 
often accepted where conscious processes are not involved. Such ac- 
ceptance does not entail a conscious choice between acceptance and re- 
jection. Rather, it is reflective of unconscious, habitual processes. 


The use of authority is one means of affecting the behavior of another. 
The subordinate who accepts an exercise of authority does not critically 
evaluate the behavior alternatives underlying the decision of his superior. 
He accepts the decision and permits it directly to affect his behavior. But 

Managerial Decision-making 9 1 

there is another means by which one individual can affect the behavior 
of another. In this case the latter individual is free to make those decisions 
which directly affect his own behavior. But, since he never has complete 
knowledge with respect to all relevant behavior alternatives and to all the 
consequences related thereto and since the ends toward which he directs 
his behavior are subject to change, it is possible for another individual or 
for others to provide him with information which can affect his decisions. 
This additional information simply adds to or changes the relevant factors 
(means and ends) which he otherwise would take into account in arriving 
at his decision. It might or might not result in a decision different from the 
one that would otherwise be made. In any event, the individual, taking 
the additional information into account, freely arrives at his own de- 
cision. Such provision of relevant information by one person to another 
(who then takes that information into account in arriving at a decision) 
will be called "influence." 7 The individual who exercises influence may 
offer advice, make suggestions, enter into discussions, persuade, use propa- 
ganda, and the like; but he does not exercise authority. In so doing, he 
indirectly affects the behavior of another. 

Managers make decisions to affect, both directly (through authority) 
and indirectly (through influence), the behavior of their subordinates. 
And, likewise, others make decisions which similarly affect the behavior 
of managers. In the two sections to follow the next one, the implications 
of these decisions for the performance of the functions of managers will 
be considered in detail. But, first, some additional factors relevant to 
decision-making will be explored. 


With respect to any given problem involving the necessity of coming 
to a decision, there are typically many desirable behavior alternatives 
from among which a choice might be made. But, for reasons to be dis- 
cussed below, the individual who must make the decision is not always 
free to choose from among all these desirable behavior alternatives. Some 
of them may be excluded, by one means or another, from his range of 
choice. It is only from among those alternatives which remain-the avail- 
able alternatives— that a choice may be made. 

In the discussion which follows, it will be said that an individual ex- 
ercises discretion with respect to available alternatives, since discretion is 
the power of free decision, of undirected choice. And the available al- 
ternatives pertinent to any given decision will be considered as falling 

7 No available term is completely satisfactory to connote the meaning here in- 
tended. The term used seems most closely to conform to that meaning. 

92 Decision-making and the Administrator 

within a sphere of discretion. For each problem calling for decision, there 
is such a sphere. A sphere of discretion has limits within which the ex- 
ercise of discretion is confined. Those factors which set the limits to 
spheres of discretion— which restrict, restrain, or limit the exercise of dis- 
cretion to available alternatives— will be referred to as "constraints." The 
types of constraints which thus define spheres of discretion are numerous 
and call for more detailed attention. 

i. Authoritative constraints.— A subordinate may have designated for 
him by his superior certain behavior alternatives which cannot be con- 
sidered by him in the making of a given decision. Thus a salesman might 
be told by his superiors that all sales made of a given item must be made 
at a price falling within a specified range. Within that range the salesman 
may exercise discretion. Of all the constraints, the authoritative is the only 
one that is personal in nature— that is imposed by one or more individuals 
on another. 

2. Biological constraints.- -When a decision is being made which will 
directly affect the behavior of the individual making the decision or the 
behavior of another, the sphere of discretion of the decision-maker can 
be constrained by certain biological characteristics of himself or of the 
other individual, as the case may be. These characteristics may be perma- 
nent in nature (a human being cannot fly), or they may be temporary 
and therefore subject to change (a person may not now know how to 
operate a lathe, but he may be able to learn to do so). 

5. Physical constraints.- -The constraints of the physical environment 
are ever present. They include such factors as geography, climate, physi- 
cal resources, man-made objects, the chemical elements, as well as physical 
and chemical laws. These factors are typically important in defining 
spheres of discretion. 

4. Technological constraints.— These constraints are determined by the 
state of the arts. For example, in determining how to make a given prod- 
uct, the decision-maker is limited in his choice to those alternatives which 
are technologically possible. 8 

5. Economic constraints- -In a freely competitive economic system, 
prices of products and of productive services are impersonally determined 
through the operation of market forces. To the individual or business 
enterprise in the system, these prices are "givens." The same is true of 
consumer wants. These "givens" are constraints with respect to economic 
decisions relating to maximization. Furthermore, the economic resources 
available to an individual or an enterprise are often also important eco- 
nomic constraints in decision-making. 

These types of constraints, where relevant, define spheres of discretion. 
They determine those behavior alternatives which are not available for 

8 Of course, he may attempt to extend the state of the arts, but such an extension 
involves a different kind of problem. 

Managerial Decision-making 93 

choice. It is from among those alternatives which remain— the available 
alternatives— that a choice is made. Decision-making, then, is judgment 
exercised within constraints. 


In the preceding section dealing with authority, it was stated that au- 
thority is used by a superior directly to affect the behavior of a subordi- 
nate. The ways in which a superior can so affect the behavior of a sub- 
ordinate can now be considered. 

1. The superior can impose constraints on a sphere of discretion of a 
subordinate as discussed above, thereby limiting the subordinate's discre- 
tion to the behavior alternatives which remain. 

2 . The superior can completely eliminate spheres of discretion from the 
province of a subordinate. In this case the subordinate is permitted no 
discretion with respect to the given problems, and no behavior is expected 
of him with respect to them. 

3. The superior can impose a decision on the subordinate to the effect 
that the subordinate act in a particular manner. Here, again, no discretion 
with respect to the problem is permitted to the subordinate, but specified 
behavior (including forbearance) is expected of him. 

Each of these devices which might be used by a superior stems from a 
decision made by him and results in some direct effect upon his sub- 
ordinate's behavior. 



It has previously been pointed out that managers make decisions to af- 
fect, both directly (through authority) and indirectly (through influ- 
ence), the behavior of their subordinates. In this section consideration 
will be given to the implications of these managerial decisions to the 
behavior of the subordinates— the managed. 

It has been seen that the relatively isolated individual is faced with in- 
surmountable difficulties in his attempt to achieve rational behavior. But 
when individuals become members of organized groups, it is at least 
possible for their behavior to achieve a high degree of rationality when 
viewed in terms of group purposes. The decisions of managers, operating 
upon the managed through authority and influence, make this possible in 
ways which will be examined below. 

In the discussion which follows, it will be important to remember that 
the term "managed" includes most managers and all nonmanagers. A 

94 Decision-making and the Administrator 

manager is such only with respect to his subordinates; he is managed with 
respect to his superior. Of all the managers in an enterprise, only the head 
of the supreme complex is a manager who is not at the same time being 
managed. The concern in this section is with the managed as such, includ- 
ing individuals who are both managers and nonmanagers. 


The decisions of managers (made to organize, direct, or control respon- 
sible subordinates) operate to increase the rationality of the behavior of 
subordinates— when viewed in terms of enterprise purposes— in the follow- 
ing ways. 

i. Decisions are made by superiors which define enterprise purpose. 
This purpose is the end for the attainment of which the specialized services 
of the members of the group are being contributed. It is important that 
the decisions made by each member of the group be made with reference 
to the group end and not a differing personal end. Through training it is 
possible to indoctrinate individuals in the enterprise purpose; through in- 
centives, to induce individuals to accept it; and through supervision, to 
insure that the enterprise purpose will guide individual decisions. 

2. Superiors establish the criterion of rationality to guide subordinates 
in making the choices which they are called upon to make. It will be re- 
membered that this criterion requires that a choice be made between al- 
ternatives which will maximize results (the degree of attainment of the 
relevant end) at a given cost. For a business firm seeking to maximize 
profits, this cost is a money cost. As in the preceding case, training, incen- 
tives, and supervision are the relevant managerial devices to be used in 
establishing the criterion of rationality as the basis for individual choice. 

3. In establishing the degree and type of specialization to be effectuated 
within an enterprise, superiors thereby define the general kind of activity 
to be expected of individuals filling particular positions. Such definition 
significantly reduces the number of spheres of discretion which are rele- 
vant to the particular activity to which an individual is assigned. This 
limitation is an aspect of the managerial function of organization. 

4. Another relevant aspect of the function of organization is the deter- 
mination of lines of formal authority. This determination establishes for 
the subordinate the individual (or individuals) to whom he is to look for 
decisions made to affect his behavior. 

5. With respect to those spheres of discretion relating to the general 
kind of activity expected of a subordinate, superiors frequently impose 
additional constraints, thereby limiting the number of available behavior 
alternatives from among which the subordinate is expected to choose. 

Managerial Decision-making 95 

6. Superiors can provide subordinates with relevant information. This 
information may relate to behavior alternatives about which the subordi- 
nate is not aware. Or it may relate to the consequences attendant upon 
specific behavior alternatives. This information may be supplied through 
training, through the use of reports and memoranda, through conversa- 
tion, and the like. 

7. Superiors may request that particular decisions be made at or by a 
specified time. Such requests are stimuli which direct the attention of sub- 
ordinates to designated problems and therefore initiate the decision-mak- 
ing processes at particular moments of time. 

8. With respect to given problem areas, superiors may expect specific 
behavior responses of their subordinates which permit no discretion to 
them. Here the subordinate is not expected to make a decision to guide his 
own behavior but simply to act in the manner specified by his superior. 
The superior may specify the action of the subordinate through an on-the- 
spot order; or he may use such devices as rules, regulations, routines, 
standing orders, policies, standard methods and procedures, and the like 
to accomplish the same purpose. In this connection it should be noted that 
the same purpose can also be accomplished by selecting people with de- 
sired attributes or by training them. Through selection or training, par- 
ticular individual modes of response can reasonably be insured so that the 
number of direct orders which must be given can be reduced. Thus, if a 
novice is hired to do clerical work, his superior must specifically tell him 
what to do, how to do it, etc. But if a trained person is hired or the novice 
is trained, then such specific orders are no longer necessary. 

In the summary, the subordinate (the individual managed) is expected 
to focus his attention on a greatly restricted number of problems calling 
for decision. With respect to these problems, authoritative constraints are 
often imposed on the pertinent spheres of discretion which further limit 
his range of choice. Information is provided which calls the subordinate's 
attention to behavior alternatives relevant to particular decisions and 
which adds to his knowledge of the consequences attendant upon those 
behavior alternatives under consideration. The ends toward which his 
decisions must be directed are specified for him, as is the criterion of ra- 
tionality to guide the choices which he must make. Lines of formal 
authority are specified for him which designate the individual (or individ- 
uals) to whom he is to look for decisions to affect his behavior. His su- 
perior often determines for him the particular problems calling for his 
decision to which he should direct his attention at specified moments of 
time. The superior often expects specific behavior responses of the sub- 
ordinate which permit no discretion to him. And through incentives and 
supervision the superior reasonably insures that all the behavior responses 
of the subordinate are in conformance with those desired. As a result of 

96 Decision-making and the Administrator 

these factors which originate with superiors, it is possible for the behavior 
of the subordinate to achieve a high degree of rationality when viewed in 
terms of enterprise purposes. 



The decisions of managers are made to affect the behavior of responsible 
subordinates in the ways considered in the preceding section. But these 
decisions are themselves not made in a vacuum. They are subject to all the 
restrictions which have previously been discussed and to influence. In 
this section, particular attention will be given to the sources from which 
stem the authority and influence which can directly and indirectly affect 
the managerial decision-making process. 


In making decisions, managers can be subject to the authority of many 
individuals and groups. The determination as to whether they will be 
subject to such authority, of course, always rests with the managers. They, 
like others, can either accept or reject any exercise of authority. If they 
accept the exercise of authority (because of positive inducements or co- 
ercion), they thereby assume a role of subordination with respect to the 
individuals or group possessing the authority. The superior can directly 
affect the behavior of the subordinate in the ways previously discussed— 
by imposing constraints on spheres of discretion, by completely eliminat- 
ing spheres of discretion from the province of the subordinate, and by im- 
posing a decision on the subordinate to the effect that the subordinate act 
in a particular manner. 

There are many individuals and groups who do, at varying times, ex- 
ercise authority with respect to managers. While no attempt will be made 
in the discussion which follows to consider all those who might authorita- 
tively impinge upon managers, the principal ones will be given attention. 

First, nearly all managers, as managed, are subject to the authority of 
their managerial superiors. This relationship was probed in the preceding 
section. At this point it is simply necessary to point out that this exercise 
of authority directly affects, among other things, the decisions made by 
the subordinate manager to affect the behavior of his own subordinates. 

Second, managers are subject to the authority of individuals who, from 
the formal point of view, are their own subordinates. At first glance, this 
may be difficult to visualize, but it is a fact the understanding of which 
is crucial to the effective performance of the functions of management. 

Managerial Decision-making 97 

It has previously been stated that the sphere of authority possessed by 
a superior is defined for, him by the sphere of acceptance of his subordi- 
nates. It is likewise true that the sphere of nonacceptance of authority of 
formal subordinates defines the sphere of nonauthority of the formal 
superior. This limiting effect imposed upon the managerial decision-mak- 
ing process by formal subordinates is indeed real. 

One of the arts of leadership is that of widening the sphere of acceptance 
of formal subordinates and, therefore, the sphere of authority of the 

Third, managers are subject to the authority of individuals and groups 
who are not members of the formal organization of the enterprise. Among 
these are the following: 

/. Governmental agencies: local, state, and -federal— Government agen- 
cies impinge upon the decision-making processes of management through 
the adoption of constitutions or charters, the passage of legislation, the in- 
terpretation of legislation by the courts, and the action of administrative 
bodies. They establish the rules of the game (the institutional framework 
within which enterprises operate), impose restrictions, demand specific 
action, settle disputes, and approve certain managerial decisions before 
these can become effective. 

2, Parties to contracts with management —When management enters 
into a contract with another party (an act involving the acceptance of the 
authority of another), it thereby agrees to meet certain obligations or to 
accept certain restrictions upon its activities. 

3, Monopolistic and monopsonistic economic groups— In those areas of 
economic activity where conditions of perfect competition do not exist, 
buyers of the enterprise's products and sellers to the enterprise of produc- 
tive services are often able, through the use of monopsonistic and monop- 
olistic power, respectively, directly to affect the behavior of managers. 
Among the monopsonists are consumer's organizations and large private 
buyers of the products of the enterprise. Among the monopolists are 
large suppliers of capital funds (banks, bondholders, etc.), raw materials, 
and labor services (unions). 9 

Because of the growing importance of unions in relation to managers, 
it is desirable to give additional attention to these monopolistic groups. 
The kinds of restrictions and the demands for particular actions which 
they impose upon managers are numerous. And they often impose these 
restrictions and demands through the threat or use of coercive economic 
power. It has previously been stated that managers are subject to the au- 
thority of individuals who, from the formal point o^ view, are their own 

9 I recognize that these monopolists and monopsonists are often parties to contracts 
with management. However, managers also often become subject to their authority 
where contractual relationships are not involved. 

gg Decision-making and the Administrator 

subordinates. In this case the limitations to the exercise of managerial 
authority are imposed by isolated individuals and informal groups. These 
same individuals, by joining together in unions, can impose the limitations 
much more effectively because of the coercive power available to strong, 
formal groups. Finally, in this connection, it should be pointed out that, 
although coercion is an important device of the union in obtaining man- 
agerial acceptance of its exercise of authority, positive inducements are 
also used. For example, a union may offer something of value to managers 
in return for an accepted limitation of managerial authority. 

4. Arbitrators.- When managers accept arbitration as a means for the 
settlement of a dispute (labor or otherwise) to which the enterprise is a 
party, they thereby assume a role of subordination with respect to the 

j. Cartels, trade-associations, and other business associations. -Enter- 
prises often are members of one or more such business associations. De- 
cisions are frequently made in these associations which are accepted as 
being authoritative by the managers of the member enterprises. 

6. The general social order.- The decisions of managers are always sub- 
ject to the general social order. Custom, tradition, convention, mores, and 
the like are the relevant authoritative principles; and sanctions (both posi- 
tive and negative) are the factors which determine managerial acceptance 
or rejection of the authority. 

Authority exercised by these individuals and groups, external to the 
enterprise, is always an extremely important factor in the direct determina- 
tion of the behavior of the managers of the enterprise. 


In making decisions, managers can also be indirectly affected by the in- 
fluence of many individuals and groups. The information supplied by 
these individuals and groups can add to or change the relevant factors 
(means or ends) which managers otherwise would take into account in ar- 
riving at decisions. It must be remembered that the individual or group 
which exercises influence may offer advice, make suggestions, enter into 
discussions, use propaganda and the like; but they do not thereby exercise 

As in the case of authority, there are individuals who are members of 
the formal organization of the enterprise who exercise influence with re- 
spect to managers. First, nearly all managers, as managed, are subject to 
the influence of their managerial superiors. This relationship, also, was 
probed in the preceding section. What is important here is that this exer- 
cise of influence indirectly affects, among other things, the decisions made 
by the subordinate manager to affect the behavior of his own subordi- 
nates. Second, managers are subject to the influence of their subordinates. 

Managerial Decision-making 99 

In this connection, it is useful to see that a manager can subdivide the 
work related to the decisions which he must make, holding subordinates 
responsible for the making of preliminary decisions (involving recom- 
mendations to him) which enter into his making of the final decision. In 
this manner the manager can be provided with information with respect 
both to relevant behavior alternatives and to the consequences attendant 
upon specific behavior alternatives. The organizational device of the staff 
is a specialized unit whose function is to provide information to the man- 
ager to whom it is attached. 10 Through the channels of upward communi- 
cation and such devices as suggestion systems, information also flows to 

Many individuals and groups who are not members of the formal or- 
ganization of the enterprise also exercise influence with respect to man- 
agers. These include, among others, governmental agencies; suppliers of 
productive services; customers; cartels, trade-associations, and other busi- 
ness associations; and consultants (including accountants, lawyers, engi- 
neers, and similar specialists). Few decisions are made by managers which 
are not indirectly affected by the influence of such individuals and groups. 


It was concluded that managers are those who use formal authority to 
organize, direct, or control responsible subordinates in order that all serv- 
ice contributions be co-ordinated in the attainment of an enterprise pur- 
pose. The use of formal authority for these purposes is the essence of 

The practice is currently prevalent on the part of many writers and 
speakers to refer to individuals and groups (external to the formal or- 
ganization of the enterprise), who exercise authority and influence with 
respect to the managers of the enterprise, as participating or sharing in 
management. These individuals and groups do not participate or share in 
management, and they are not managers (except, perhaps, of the enter- 
prises to which they belong) . They regulate managers (through authority 

10 Lewis C. Sorrell has commented upon this fact, as follows: "While much con- 
fusion obtains regarding the proper functioning of a staff in business concerns . . . , 
essentially it is coming to mean one or more persons charged with the responsibility 
of ascertaining the important trends internal and external to the business, determining 
their incidence upon the enterprise or some important division thereof, observing the 
methods employed by others for coping with similar situations, selecting the several 
alternatives of policy that appear to be most practical under all the circumstances, and 
presenting all this to the proper executives-but without authority to direct or order 
anything into execution" ("The Role of Management in the Organization of Re- 
sources for Production," Management's Adjustment to the Changing National Econ- 
omy, ed. William N. Mitchell [Chicago: University of Chicago Press, 1942], p. 37). 

i oo Decision-making and the Administrator 

and influence), but they do not manage. 11 In the present writer's view, if 
one attempted a definition of the manager broad enough to include these 
individuals and groups (the regulators), the result would be a concept of 
the manager which would be so lacking in content and sharpness as to be 
of little theoretical or practical value. 

The point is simply this: Managers head groups (either of other man- 
agers or of nonmanagers) in formal systems of co-ordination. They use 
formal authority to manage responsible subordinates. The individuals 
comprising each group are guided by an enterprise purpose, and their 
specialized service contributions are co-ordinated by their manager in the 
attainment of that purpose. On the other hand, the regulators, who are 
external to the formal organization of the enterprise, are not a part of the 
formal system of co-ordination. They use authority (but not formal au- 
thority) and influence to regulate (i.e., directly and indirectly to affect) 
the behavior of others. They and the managers are not guided by the same 
purpose, nor is co-ordination of specialized service-contributions toward 
the attainment of an enterprise purpose involved in the relationship be- 
tween them. 

To differentiate clearly between managers and regulators is to provide 
an extremely useful frame of reference for dealing with many problems, 
including those of public policy. To fail to make the differentiation can 
lead to confusion and misunderstanding, the effects of which, in this 
writer's judgment, can only be an incorrect approach to those problems. 


Richard M. Cyert, Herbert A. Simon, and Donald B. Trow* 

The authors have dissected an actual major business decision for the 
purposes of evaluating traditional assumptions regarding business decision- 
making. The example chosen concerned itself with the feasibility of using 
electronic data-processing equipment and is omitted in order to emphasize 
the conclusions reached by the authors. 

Decision-making— choosing one course of action rather than another, 
finding an appropriate solution to a new problem posed by a changing 
world— is commonly asserted to be the heart of executive activity in busi- 

11 The term "control" would be preferable, in this connection, to the term "reg- 
ulate." I do not here use the former term, since it has been used above with another 

* From Journal of Business, XXIX, 4 (1956), 237-48. Reprinted by permission of 
the University of Chicago Press. 

Observation of a Business Decision i o i 

ness. If this is so, a realistic description and theory of the decision-making 
process are of central importance to business administration and organiza- 
tion theory. Moreover, it is extremely doubtful whether the only con- 
siderable body of decision-making theory that has been available in the 
past— that provided by economics— does in fact provide a realistic account 
of decision-making in large organizations operating in a complex world. 

In economics and statistics the rational choice process is described some- 
what as follows: 

i. An individual is confronted with a number of different, specified 
alternative courses of action. 

2. To each of these alternatives is attached a set of consequences that 
will ensue if that alternative is chosen. 

3. The individual has a system of preferences or "utilities" that permit 
him to rank all sets of consequences according to preference and to choose 
that alternative that has the preferred consequences. In the case of business 
decisions the criterion for ranking is generally assumed to be profit. 

If we try to use this framework to describe how real human beings 
go about making choices in a real world, we soon recognize that we need 
to incorporate in our description of the choice process several elements 
that are missing from the economic model: 

1. The alternatives are not usually "given" but must be sought, and 
hence it is necessary to include the search for alternatives as an important 
part of the process. 

2. The information as to what consequences are attached to which al- 
ternatives is seldom a "given," but, instead, the search for consequences 
is another important segment of the decision-making task. 

3. The comparisons among alternatives are not usually made in terms 
of simple, single criterion like profit. One reason is that there are often 
important consequences that are so intangible as to make an evaluation in 
terms of profit difficult or impossible. In place of searching for the "best" 
alternative, the decision-maker is usually concerned with finding a satis- 
factory alternative— one that will attain a specified goal and at the same 
time satisfy a number of auxiliary conditions. 

4. Often, in the real world, the problem itself is not a "given," but, in- 
stead, searching for significant problems to which organizational attention 
should be turned becomes an important organizational task. 

Decisions in organizations vary widely with respect to the extent to 
which the decision-making process is programmed. At one extreme we 
have repetitive, well-defined problems (e.g., quality control or production 
lot-size problems) involving tangible considerations, to which the eco- 
nomic models that call for finding the best among a set of pre-established 
alternatives can be applied rather literally. In contrast to these highly pro- 
grammed and usually rather detailed decisions are problems of a non- 

1 02 Decision-making and the Administrator 

repetitive sort, often involving basic long-range questions about the whole 
strategy of the firm or some part of it, arising initially in a highly un- 
structured form and requiring a great deal of the kinds of search processes 
listed above. In this whole continuum, from great specificity and repetition 
to extreme vagueness and uniqueness, we will call decisions that lie toward 
the former extreme programmed, and those lying toward the latter end 
non-programmed. This simple dichotomy is just a shorthand for the range 
of possibilities we have indicated. 

It is our aim in the present paper to illustrate the distinctions we have 
introduced between the traditional theory of decision, which appears ap- 
plicable only to highly programmed decision problems, and a revised 
theory, which will have to take account of the search processes and other 
information processes that are so prominent in and characteristic of non- 
programmed decision-making. We shall do this by recounting the stages 
through which an actual problem proceeded in an actual company and 
then commenting upon the significance of various items in this narrative 
for future decision-making theory. 

The decision was captured and recorded by securing the company's 
permission to have a member of the research team present as an observer 
in the company's offices on substantially a full-time basis during the most 
active phases of the decision process. The observer spent most of his time 
with the executive who had been assigned the principal responsibility for 
handling this particular problem. In addition, he had full access to the files 
for information about events that preceded his period of observation and 
also interviewed all the participants who were involved to a major degree 
in the decision. 


The decision process to be described here concerns the feasibility of us- 
ing electronic data-processing equipment in a medium size corporation 
that engages both in manufacturing and in selling through its own widely 
scattered outlets. In July, 1952, the company's controller assigned to 
Ronald Middleton, an assistant who was handling several special studies in 
the accounting department, the task of keeping abreast of electronic de- 
velopments. The controller, and other accounting executives, thought 
that some of the current developments in electronic equipment might have 
application to the company's accounting processes. He gave Middleton the 
task of investigation, because the latter had a good background for under- 
standing the technical aspects of computers. 1 

1 The "recounting of the stages" of the problem follows this point in the article, 
but it is omitted here in order to conserve space. The discussion which follows is 
relevant generally and is not dependent on the actual case illustration. 

Observation of a Business Decision i o 3 


From this narrative, or more specifically from the actual data on which 
the narrative is based, one can list chronologically the various activities of 
which the decision process is composed. If we wish to describe a program 
for making a decision of this kind, each of these activities might be taken 
as one of the steps of the program. If the rules that determine when action 
would switch from one program step to another were specified, and if the 
program steps were described in enough detail, it would be possible to 
replicate the decision process. 

The program steps taken together define in retrospect, then, a program 
for an originally unprogrammed decision. The program would be an in- 
efficient one because it would contain all the false starts and blind alleys of 
the original process, and some of these could presumably be avoided if the 
process were repeated. However, describing the process that took place in 
terms of such a program is a useful way of organizing the data for pur- 
poses of analysis. 

In order to make very specific what is meant here by a "program," 
Chart I has been prepared to show the broad outlines of the actual pro- 
gram for the first stages of the decision process (through the first seven 
paragraphs of the narrative). 



Keeping-up program (paragraphs 1 and 2 of narrative): 
Search for and correspond with experts; 
Discuss with salesmen and with equipment users; 
Search for and read journals; 

Procurement program (paragraph 3): 

Discuss applications study with salesmen who propose it; 
Choice: accept or reject proposed study; 
(If accepted) transfer control to salesmen; 
Choice: accept or reject applications proposal; 
(If rejected) switch to consultant program; 

Consultant program (paragraphs 4 through 7): 
Search for consultants; 
Choice: best consultant of several; 
Transfer control to chosen consultant; 
Choice: accept or reject proposal; 
(If accepted): begin double-check routine; 
Request expenditure of funds; 
(If authorized) transfer control to consultants; 
And so on. 

! 04 Decision-making and the Administrator 


The various program steps of the decision process fall into several sub- 
programs, some of which have been indicated in Chart I. These subpro- 
grams are ways of organizing the activities post factum, and in Chart I 
the organizing principle is the method of approach taken by the company 
to the total problem. It remains a question as to whether this organizing 
principle will be useful in all cases. As in the present example, these 
subprograms may sometimes be false starts, but these must be regarded as 
parts of the total program, for they may contribute information for later 
use, and their outcomes determine the switching of activity to new sub- 

In this particular case the reasons for switching from one subprogram 
to another were either the proved inadequacy of the first one or a re- 
definition of the problem. Other reasons for switching can be imagined, 
and a complete theory of the decision process will have to specify the 
conditions under which the switch from one line of attack to another will 


In the whole decision-making program there are certain steps or 
"routines" that recur within several of the subprograms; they represent 
the basic activities of which the whole decision process is composed. For 
purposes of discussion we have classified these common processes in two 
categories: the first comprises processes relating to the communication 
requirements of the organization; the second comprises processes relating 
directly to the solution of the decisional problem. 


Organizational decision-making requires a variety of communication 
activities that are absent when a decision is made in a single human head. 
If we had written out the program steps in greater detail, many more in- 
stances of contacts among different members of the organization would 
be recorded than are now explicit in the narrative. The contacts may be 
oral or written. Oral contacts are used for such purposes as giving orders, 
transmitting information, obtaining approval or criticism of proposed ac- 
tion; written communications generally take the form of memorandums 
having the purpose of transmitting information or proposing action. 

The information-transmitting function is crucial to organizational de- 
cision-making, for it almost always involves acts of selection or "filtering" 
by the information source. In the present instance, which is rather typical 

Observation of a Business Decision 1 05 

in this respect, the consultants and subordinate executives are principal 
information sources; and the controller and other top executives must 
depend upon them for most of their technical information. Hence, the 
subordinate acts as an information -filter and in this way secures a large 
influence over the decisions the superior can and does reach. 

The influence of the information source over communications is partly 
controlled by checking processes— for example, retaining an independent 
expert to check consultants which give the recipient an independent in- 
formation source. This reduces, but by no means eliminates, filtering. The 
great differences in the amounts and kinds of information available to the 
various participants in the decision process described here emphasize the 
significance of filtering. It will be important to determine the relationship 
of the characteristics of the information to the resultant information 
change and to explore the effects of personal relations between people on 
the filtering process and hence upon the transmission of information. 


Alongside the organizational communication processes, we find in the 
narrative a number of important processes directed toward the decision 
problem itself. One of the most prominent of these is the search for 
alternative courses of action. The first activities recounted in the narra- 
tive—writing letters, reading journals, and so on— were attempts to discover 
possible action alternatives. At subsequent points in the process searches 
were conducted to obtain lists of qualified consultants and experts. In 
addition to these, there were numerous searches— most of them only im- 
plicit in the condensed narrative— to find action alternatives that would 
overcome specific difficulties that emerged as detail was added to the 
broader alternatives. 

The data support strongly the assertion made in the introduction that 
searches for alternative courses of action constitute a significant part of 
non-programmed decision-making— a part that is neglected by the classical 
theory of rational choice. In the present case the only alternatives that 
became available to the company without the expenditure of time and 
effort were the systems proposals made early in the process by representa- 
tives of two equipment companies, and these were both rejected. An im- 
portant reason for the prominent role of search in the decision process is 
that the "problem" to be solved was in fact a whole series of "nested" 
problems, each alternative solution to a problem at one level leading to a 
new set of problems at the next level. In addition, the process of solving 
the substantive problems created many procedural problems for the or- 
ganization: allocating time and work, planning agendas and report presen- 
tations, and so on. 

! 6 Decision-making and the Administrator 

Examination of the narrative shows that there is a rich variety of search 
processes. Many questions remain to be answered as to what determines 
the particular character of the search at a particular stage in the decision 
process: the possible differences between searches for procedural alterna- 
tives, on the one hand, and for substantive alternatives, on the other; the 
factors that determine how many alternatives will be sought before a 
choice is made; the conditions under which an alternative that has tenta- 
tively been chosen will be subjected to further check; the general types of 
search strategies. 

The neglect of the search for alternatives in the classical theory of deci- 
sion would be inconsequential if the search were so extensive that most of 
the alternatives available "in principle" were generally discovered and 
considered. In that case the search process would have no influence upon 
the alternative finally selected for action. The narrative suggests that this 
is very far from the truth-that, in fact, the search for alternatives termi- 
nates when a satisfactory solution has been discovered even though the 
field of possibilities has not been exhausted. Hence, we have reason to 
suppose that changes in the search process or its outcome will actually 
have major effects on the final decision. 

A second class of common processes encompasses information-gathering 
and similar activity aimed at determining the consequences of each of 
several alternatives. In many decisions, certainly in the one we observed, 
these activities account for the largest share of man-hours, and it is 
through them that subproblems are discovered. The narrative suggests 
that there is an inverse relation between the cost or difficulty of this in- 
vestigational task and the number of alternative courses of action that are 
examined carefully. Further work will be needed to determine if this rela- 
tion holds up in a broader range of situations. The record also raises nu- 
merous questions about the kinds of consequences that are examined most 
closely or at all and about the conditions under which selection of criteria 
for choice is prior to, or subsequent to, the examination of consequences. 
Another set of common processes are those concerned with the choices 
among alternatives. Such processes appear at many points in the narrative: 
the selection of a particular consulting firm from a list, the choice between 
centralized and decentralized electronic-data-processing systems, as well 
as numerous more detailed choices. These are the processes most closely 
allied to the classical theory of choice, but even here it is notable that 
traditional kinds of "maximizing" procedures appear only rarely. 

In some situations the choice is between competing alternatives, but in 
many others it is one of acceptance or rejection of a single course of ac- 
tion-really a choice between doing something at this time and doing 
nothing. The first such occasion was the decision by the controller to 

Observation of a Business Decision 1 07 

assign Middleton to the task of watching developments in electronics, a 
decision that initiated the whole sequence of later choices. In decisions of 
this type the consequences of the single alternative are judged against 
some kind of explicit or implicit "level of aspiration"— perhaps expressed 
in terms of an amount of improvement over the existing situation— while 
in the multiple -alternative situations, the consequences of the several al- 
ternatives are compared with each other. This observation raises a host of 
new questions relating to the circumstances under which the decision will 
be formulated in terms of the one or the other of these frameworks and 
the personal and organizational factors that determine the aspiration levels 
that will be applied in the one-alternative case. 

Another observation derivable from our data— though it is not obvious 
from the condensed narrative given here— is that comparability and non- 
comparability of the criteria of choice affects the decision processes in 
significant ways. For one thing, the criteria are not the same from one 
choice to another: one choice may be made on the basis of relative costs 
and savings, while the next may be based entirely on non-monetary cri- 
teria. Further, few, if any, of the choices were based on a single criterion. 
Middleton and the others recognized and struggled with this problem of 
comparing consequences that were sometimes measured in different, and 
incomparable units, and even more often involved completely intangible 
considerations. The narrative raises, but does not answer, the question of 
how choices are made in the face of these incommensurabilities and the 
degree to which tangible considerations are overemphasized or under- 
emphasized as compared with intangibles as a result. 


We do not wish to try to transform one swallow into a summer by 
generalizing too far from a single example of a decision process. We have 
tried to illustrate, however, using a large relatively non-programmed de- 
cision in a business firm, some of the processes that are involved in business 
decision-making and to indicate the sort of theory of the choice mecha- 
nism that is needed to accommodate these processes. Our illustration sug- 
gests that search processes and information-gathering processes constitute 
significant parts of decision-making and must be incorporated in a theory 
of decision if it is to be adequate. While the framework employed here- 
and particularly the analysis of a decision in terms of a hierarchical struc- 
ture of programs— is far from a complete or finished theory, it appears to 
provide a useful technique of analysis for researchers interested in the 
theory of decision as well as for business executives who may wish to re- 
view the decision-making procedures of their own companies. 

1 08 Decision-making and the Administrator 


R. K. Gaumnitz and O. H. Brownlee* 

The authors take note of the impact of the electronic computer on data 
manipulation and the increasing application of mathematics to administra- 
tive problems involving decisions. The authors draw the conclusion that 
the decision-maker does not have to be a mathematician to make effective 
use of mathematical techniques or electronic computers, but they also note 
that the resulting lack of easy communication between the decision-maker 
and the mathematician may make co-operation difficult and even haz- 

Business executives recently have been flooded with descriptions of the 
wonders that can be performed with the aid of such devices or techniques 
as electronic computers, mathematical programing (linear programing and 
some aspects of operations research, for example), and servomechanisms 
or information theory. Some of these are the result of refinements in 
devices and techniques that have been used or known for a relatively long 
time. Others are relatively new. All of them require fairly extensive 
knowledge of certain branches of mathematics, on the part of someone, 
in order that they can be thoroughly understood and operated to ad- 

Does this mean that mastery of calculus, matrix algebra, probability 
theory, and topology is becoming a management prerequisite? Definitely 
no. Does it mean that the top executives of tomorrow must be better 
skilled in fundamental operations" like calculating percentages and work- 
ing out equations? Not necessarily. Does it mean that more understanding 
of the mathematical approach to problems will be useful to the man who 
makes important business decisions? Unqualifiedly yes. 

The recent developments in this field do not change the basic character 
of the decision making that the businessman must do. Rather, they permit 
more rigorous formulation of the problems he faces, facilitate the figuring 
of possible situations for him to consider, and add explicitness and quanti- 
fication to the terms in which he frames his decisions. At the same time, 
it cannot be denied that a top executive would benefit from knowledge of 
certain kinds of mathematics. This has always been true. The only differ- 
ence is that now, because of the technical advances that have been made, 
the task of understanding and communicating is more difficult and the 
potential of usefulness is richer. 

*From Harvard Business Review, XXXIV, 3 (1956), 48-56. Reprinted by permis- 
sion of the Harvard Business Review. 

Mathematics for Decision-makers 1 00 

The real question of the moment is how much knowledge the executive 
should have in view of (a) what it can contribute and (b) what is en- 
tailed in acquiring it. 


Some of the new developments have been responsible for considerable 
speculation and comment about the importance of mathematics to man- 
agement. How much difference will they actually make as their use be- 
comes more and more widespread? 


One may distinguish two general types of situations involving electronic 
computers: (i) the transfer to computers of work now performed by 
clerks or by lower order machines (i.e., machines that are slower and lack 
"memories"); and (2) the use of modern machines or methods to carry 
out clerical jobs not previously undertaken at all because of the difficulty 
of doing so by conventional methods. 

In the first instance, all that the computer or the data-processing device 
does is to perform routine clerical operations more efficiently than has 
been possible without the aid of such a machine. Thus, for instance, com- 
puters and other electronic equipment may process certain insurance or 
public utility billing data much less expensively and/or at greater speed 
than is possible with traditional methods and machines. 

Here it is clear that the top executive, at least, needs no special training 
in mathematics to benefit from the use of such equipment. While some- 
one in the organization surely will have to know a great deal about how 
to instruct and operate the electronic equipment, there is no more reason 
for the executive to be informed of the details of this than for him to 
know how to operate a new folding machine that will speed mass mailing. 
These are matters for technicians, not executives. 

In the second instance, some problems that are not solved at present, 
simply because the time required (or the cost involved) is too great if 
traditional equipment and techniques are used, become economically fea- 
sible by using computers and other electronic equipment. 

Substantial reductions in cost are sometimes more important than the 
time-saving illustrated above. Thus, if clerks cost only $20 per year, cer- 
tain sales analyses would be performed in many firms; utilizing modern 
calculating equipment may actually achieve the equivalent of this low 
price for such work. 

Here again the executive need have no special knowledge of mathe- 
matics to benefit from the new equipment. However, he knows best what 

j j Decision-making and the Administrator 

would be useful to him in making decisions; and if he has some under- 
standing of the practicability of different statistical manipulations, he will 
be in a better position to identify his needs and get them served. 


From the standpoint of the business executive, the new techniques for 
stating and solving problems are like computing machines in that they too 
make it possible to solve some problems much more readily than could 
be done before. Linear programing is one of the techniques and it can 
serve to illustrate the extent of mathematical knowledge needed by the 

No special mathematical training is required to understand the results of 
typical applications of linear programing or, for that matter, to under- 
stand what is involved in it. 

Frequently solution by linear programing methods requires, as a prac- 
tical matter, access to an electronic computer. But essentially the tech- 
nique does nothing that could not be done, given enough time and enough 
clerks, with the use of simple mathematical techniques or ordinary me- 
chanical computers. 

What the technique does is to make possible more rapid solutions of 
certain problems, particularly those where choices must be made from 
among a large number of alternative courses of action, some of which may 
be more attractive than others in terms of higher profit, lower cost, or 
some other characteristic. In other words, it simply systematizes and 
makes efficient the process of identifying the most attractive course or 
courses of action. 

Why the technique works, and how it works, is not vital information 
for the executive. What matters to him is that the technique does identify 
the best solution. Rather than being capable of performing the necessary 
computations, he needs to be able to communicate with the experts about 
the problems on which they can help him. 

Techniques also are being developed for solving selected scheduling 
problems more efficiently-for example, the order in which products 
should be manufactured with available labor and equipment. The relation 
of the executive to these techniques parallels his relationship to linear pro- 
graming, in that he requires understanding of the general approaches, 
rather than special mathematical knowledge, in order that his company 
may enjoy some of the benefits of the new procedures. 


While the executive need not understand the basic principles underlying 
an electronic computer or how to operate it, and need not know why and 

Mathematics for Decision-makers 1 1 1 

how techniques such as linear programing work, there is still a lot that he 
needs to know about mathematics— perhaps not so much more in quantity 
than in the past, but of a different kind. 

Mathematics can be viewed as a language that is particularly useful in 
formulating certain types of problems, because it focuses on only the es- 
sential factors and their logical relationships. Thus, it is a language that 
makes it easier for a person to see the implications of certain kinds of ac- 
tions. Undoubtedly there are, and will be, successful executives who know 
little or no mathematics or who never use mathematical notation in their 
thinking. But a person of given intelligence can see certain problems more 
clearly and solve them more readily with the aid of mathematics. 

Among the classes of problems that one can think about efficiently in 
the mathematical language— i.e., formulate and solve— are certain business 
problems commonly presented for executive decision. Such problems 
typically require finding an optimum— usually a maximum or a minimum— 
from among a number of possible values. Examples are the problems of 
minimizing the cost of producing a given output, choosing a collection of 
products that will maximize revenue, arriving at a price or markup for a 
product, and selecting the amount of advertising outlay for each of a num- 
ber of different media. Merely identifying a problem as one of maximum 
or minimum values frequently helps to clarify it. 

Also, the implications of certain actions can be perceived more readily 
when problems are recognized as problems of maxima and minima. For 

Some businessmen seem to be overimpressed with the fact that a large 
volume of production means a relatively low per unit overhead cost— the de- 
sirability of "spreading the overhead." 

Actually it is only the changes in total cost and total revenue that should 
determine whether output should be changed. The objective is to maximize 
the difference between total revenue and total cost. Changes in total cost are 
not affected by the level of fixed costs and therefore "overhead" items are 
irrelevant except when deciding whether to produce at all. 

The above relationships have not been universally appreciated- by business 
executives; some of this lack of perception may be due to traditional account- 
ing methods and concepts. But persons who are able to describe revenue and 
cost algebraically as they are related to output, and who possess even a vague 
familiarity with the calculus, would not be expected to make such an error. 

It frequently is desirable for an executive to examine problems in 
quantitative form. The habit of thinking in quantitative terms has been 
acquired by some people with little mathematical training. Using the 
"sharp pencil" approach to many problems often characterizes successful 
executives. Insofar as mathematical training strengthens and extends this 
approach, it will aid in making effective decisions. 

Thinking in quantitative terms is commonplace where the factors to be 
considered are expressed as numbers— wage rates, man-hours, raw mate- 
rials, costs, transportation charges, and so on— which in turn are used to 

1 1 2 Decision-making and the Administrator 

obtain total cost, total receipts, profits, or some other figure of interest. 
However, there are other situations where the variables are not so readily 
measured but where quantification is still possible and might lead to more 
reasonable decisions than those reached with only vague quantification. 
For example: 

Deciding whether to employ a more accurate but more costly price or sales 
forecasting procedure requires assigning a schedule of values to greater ac- 

Or making a sensible decision between (a) shipping an uninspected product 
and taking a chance that some customers may receive an unsatisfactory item 
and (b) shipping a product late but properly inspected, which would prob- 
ably involve an explicit evaluation of possible losses from the alternative 
courses of action. 

The outcomes of most business problems are not certain; some, how- 
ever, can be described in probability terms. One usually cannot determine 
exactly what output will be produced by a given number of men and a 
given number of machines, by what amount total sales will change as a 
result of a particular advertising campaign, or what price will be paid for 
a certain raw material or be received for a particular kind of product. 
However, it is sometimes helpful to make some kind of probability state- 
ments about such phenomena. For instance: 

"The chances are three out of four that the price will be between $8 and $9." 
"The chances are nine out of ten that the per acre corn yield will be less 

than 60 bushels." 

"We have better than a 50-50 chance of cracking the cat food market by 

this campaign." 

Some understanding of probabilities is particularly useful in making de- 
cisions based on price forecasting, quality control, or inventory control 
procedures. Without it there can be trouble. Quite a few businessmen 
think that if the chances are one out of three that a certain desired result 
will ensue, a person can be sure of achieving it in three tries. But actually 
what the statement means is that the likelihood of the event not happening 
is twice as great as the likelihood of its happening, and the implication 
for the businessman might be that he had better not undertake the risk 
unless he can withstand the cost of failure. 


Let us take for granted that the executive is not the person who per- 
forms the calculations, writes elaborate equations, expresses inequalities, 
and so on; and that he has access to people who are trained in the relevant 
mathematics— if, indeed, mathematical formulations and analytical methods 
are to be employed on any sizable scale. On this basis, there is obviously a 
problem of communication between the executive and the technician. 

Mathematics for Decision-makers i 1 3 

The extent of the difficulty here depends on the intelligence and breadth 
of knowledge of both the technician and the executive involved. If the 
technician knows little or nothing about the "practical facts" surrounding 
specific problems, he will have more difficulty getting relevant informa- 
tion from the executive than if he has some clues as to what is significant. 
Similarly, if the executive is virtually illiterate with respect to mathe- 
matical matters, he will have trouble understanding what it is the tech- 
nician is (or ought to be) after unless the latter is unusually skillful in ex- 
plaining what it is he needs: For instance: 

Some executives would understand at once what was desired if the tech- 
nician asked about the elasticity of demand for the product under considera- 
tion. The response would depend chiefly on the executive's familiarity with 
the terminology of economics. But a good many executives would not really 
understand the question if asked whether a given change in quantity produced 
a relatively greater change in price. The key word is "relatively," and it is 
quite commonly misunderstood by executives. 

This trivial point illustrates one kind of communication problem that exists— 
an unfamiliarity with the exact meaning of words in mathematics which are 
used loosely or carelessly in ordinary speech. 

The executive must at least be able to understand what the technician 
wants and needs in order to mathematize a given problem satisfactorily or 
to recognize whether the problem is one that can actually be formulated 
mathematically. Mathematics cannot make something out of nothing; an 
incorrect formulation can be just as serious as inaccurate data— perhaps 
more so, for the possibility of errors in the data can often be provided 
for. For example: 

Suppose the mathematician is asked to tackle a problem of allocating over- 
head to various machine operations. To do this intelligently he may need to 
know such seemingly (to the executive) unrelated matters as the alternative 
uses of the various machines and the different market returns of the items pro- 

Further, he may want an idea of what degree of precision is needed in the 
answer for the purpose of making decisions; for if a range of error of plus or 
minus 10% is admissible, the procedure may be one fourth as cumbersome 
and expensive as if the error were to be held to 5 % . 

Similarly, to appreciate the limitations of answers supplied by his tech- 
nicians, the executive usually cannot be completely devoid of mathe- 
matical knowledge. 

In an effort to limit the number of variables considered in the mathe- 
matical formulation of problems, the technician tries to identify the more 
important factors in any given relation. One reason for this objective is 
evident to anybody familiar with the speed with which computational 
burdens pyramid as the variables increase in number; another reason is so 
that the results may be more meaningful. Executives are inclined to be im- 
patient if the technician persistently wants to limit the number of variables 

1 14 Decision-making and the Administrator 

being actively considered. Their usual attitude is: '.'You've got to think 
about all the angles." To illustrate: 

A textile manufacturer may wish to test a proposed new dish towel for 
market reaction by trial sales in a number of stores. Perhaps he would like to 
know about size, price, and color; and he has five possible alternatives for each 
of these factors. The possible combinations amount to 125. There would have 
to be a fantastic number of stores in the sample to get any significant results— 
and there probably wouldn't be that many stores available where conditions 
could be held comparable (as to time, size, type of customers, competition, and 
so on). 

So the technician asks which factor or factors are important. In this case it 
might be that the size of the towel was limited by the width of the looms in 
the mill, and perhaps the sales records on previous towels already indicate the 
most popular colors, so the test can be narrowed primarily to the question of 

But even here the technician can be helped to help the executive if the latter 
will accept findings which are, say, "70% probable." In other words, if the 
manufacturer has enough flexibility in changing prices subsequently so that he 
can take the 30% risk of being wrong in his initial price, then the number of 
stores that must be included will be fewer in number (so the research can be 
quicker and more economical), and they can be selected more precisely for 
comparability (so the results will also be more dependable within the ac- 
ceptable limitations of what is needed for making a practical business decision). 

Part of the explanation for executives' impatience is that they apparently 
have difficulty in distinguishing degrees of importance or feel that only 
perfectly precise statements are usable. When the technician attempts to 
pin down the executive as to how variables are related or how important 
they are, a very common— and quite unhelpful— reply is: "That's hard to 
say." If the businessman does not understand what is being asked (or is 
unable or unwilling to answer the question), the possibilities of mathe- 
matical treatment are noticeably limited. 


There probably is no limit to the amount of mathematics that could be 
used by business executives. However, to acquire such knowledge would 
entail costs— in terms of other branches of knowledge that could be ac- 
quired—which make the practical limit relatively low. Examples of ap- 
plications of special branches of mathematics to business problems abound 
in the literature. 1 Fuller utilization of the mathematics already possessed 
by executives rather than expansion of mathematical knowledge might be 
a reasonable first step. 

It should be re-emphasized that the executive is not expected to be 
active in the calculation and solution stage of problems being treated 

1 In addition to the HBR "Statistical Series" there are such publications as Eco- 
nometrica, Management Science, and the Journal of the Operations Research Society 
of America. 

Mathematics for Decision-makers i 1 5 

mathematically. His work comes earlier-stating the problem with enough 
exactness so that it can be translated into mathematical language. Relevant 
factors must be identified and their relationship described; this will cer- 
tainly require the executive's close attention. The executive is likely to be 
the only one able to state the proper "weights" or relationships to be em- 
ployed in any mathematical statement of the problem. This is the key 
point in the process so far as the executive is concerned; it is here that he 
makes his unique contribution. 

In making decisions in the traditional way— by entirely subjective 
methods— the executive somehow combines his evaluations of different 
factors in a manner that leads to a decision. He must decide how a given 
factor relates to the problem and what its magnitude will be, and must 
take into account any joint effects there may be. The influence of this 
particular factor is considered along with that of all the other factors the 
executive decides to take into account; the final decision rests on the com- 
posite effect of all factors. 

The whole process being subjective, most of the variables are evidently 
not thought of in specific quantitative terms. The executive apparently 
thinks to himself, "That won't amount to much," or "This will be quite 
important," or some other evaluation of this general type. In a mathemati- 
cal formulation the technician must elicit as much of this evaluation as 
possible from the executive in order to bring it to bear on the solution. 
While it is possible to do something even with loose statements of the 
kind quoted above, much more can be done with a more specific relation- 
ship even though it may be only approximate. Thus: 

It is better to say, "It typically costs $200 to get a customer back if we lose 
him" than to say, "It costs money to get customers back." It is better to say, 
"There are four chances out of five we'll lose a customer if we don't make 
immediate deliveries nine-tenths of the time," than to say, "Customers won't 
keep coming back forever unless we deliver promptly." 

Ordinarily the mathematician cannot himself provide the relationships 
that are present; he must get these relationships described to him by man- 
agement. He may, in some instances, be able to suggest mathematical pro- 
cedures for evaluating data the company may have, and to help get certain 
relationships identified (such as those used in respect to customers in the 
last paragraph). But he can do practically nothing with respect to decid- 
ing what top management really feels the relative importance of various 
company objectives to be. To be included in mathematical models com- 
pany objectives or policies must be stated in somewhat mathematical 
terms; and only the management can typically provide the information 
necessary to do this. For example: 

In a problem of setting up an inventory control system the executive must 
indicate what level of customer service he wishes to maintain and how much 
variation in employment he considers to be tolerable. 

1 1 6 Decision-making and the Administrator 

Thus his statement of the problem might be: "What will be the most eco- 
nomical production schedule that will provide us with sufficient stock so that 
we can fill customers' orders immediately nine times out of ten and not have 
more than a 25% variation in employment?" 

A mathematical statement of the verbal explanation often facilitates the 
perception of implications of certain relationships, inconsistencies, and so 
on. Thus: 

In the above example, attention might be called to the rate of fluctuation in 
customers' orders; and if it turned out that lumping of orders around the first 
of the month made a difference of 90% in the amount of safety stock needed, 
management might be prompted to take steps to reduce the monthly peak. 

Unfortunately there is no reliable method of describing those classes of 
problems that are handled more efficiently with the aid of mathematical 
techniques. Where large numbers of variables are involved or chains of 
reasoning are long, a mathematical formulation is likely to be efficient if 
not indispensable, but even quite simple problems are sometimes best 
handled with mathematical tools. 


There are many examples in typical firms of failure to utilize even the 
most elementary mathematics in evaluating problems. Here are two glar- 
ing instances: 

Comparing the cost of achieving an outcome and the value of that outcome 
should be a regular part of examining procedures and proposals, yet frequently 
is not. There is often inadequate consideration of what it costs to prevent a 
given kind of loss as compared with the size of the loss. If it costs $10,000 to 
provide a fire protection system which at most would prevent one $5,000 fire, 
the decision to install the system would be a poor one (if there were no 
secondary effects from the $5,000 fire, like disrupting the production line). 

Accounting procedures are sometimes used which involve substantial outlays 
to develop data, or to provide accuracy in data, which simply cannot be 
justified in terms of the value of the results. A firm storing oil is reputed to 
have estimated its oil stocks by measuring the oil levels in tanks with a stick 
and then to have spent hundreds of clerk-hours in trying to reconcile to the 
pound its book inventory with that observed in the tanks. 

Accounting data prepared for utilization in decision situations are often, 
at considerable extra expense, calculated to the penny when the decision 
requires data accurate only to the nearest thousand dollars. Large numbers 
of routine papers have been subjected to "clearance" by senior officers, at 
great expense over the years, largely because a junior official once made a 
mistake that cost the company, say, $1,400. No special training in mathe- 
matics is needed in order to make obviously reasonable assumptions and 

Mathematics for Decision-makers 1 1 7 

to evaluate situations of this kind. Knowing and using arithmetic is ade- 

If any but the simplest relationships are present, or if the generality of a 
particular relationship is to be investigated, it is frequently valuable to 
utilize a more general language, namely, algebra. The use of this branch 
of mathematics greatly enlarges the scope of problems that can be effi- 
ciently treated. A further extension of the range of problems is achieved 
by using differential calculus. Here the obvious class of problems includes 
many having to do with maxima and minima, which were mentioned 
earlier. For example: 

Take a problem of determining the amount of dollars to spend for advertis- 
ing and the quantity of product to manufacture. The relationship between the 
price at which the product can be sold, the quantity placed on the market, and 
the amount spent for advertising is assumed to be known, e.g., P = f(x, a). 
Also known is the relationship between production cost and output, e.g., 
C = (f>(x). To be decided are x, the amount to produce, and a, the amount to 
spend for advertising. By differential calculus, the values for x and a that will 
maximize profit can be obtained. 

Indeed, for every kind of mathematics some business problem doubtless 
exists that can most easily be solved using the special features of the mathe- 
matics. And it is probably true that additional mathematical knowledge, 
of almost any kind, improves the mathematical "maturity" of the person. 
For most executives some knowledge of algebra, of the meaning of a 
functional relationship, and the rudiments of probability theory would 
constitute a fair beginning. For specific executive situations knowledge of 
special types of mathematics might be exceptionally valuable. But, in gen- 
eral, the higher his level, the less likely the executive is to require specific 
and technical knowledge of mathematics. 




Kenneth E. Boulding* 

Professor Kenneth E. Boulding examines the case for admitting the 
existence of an administrative science by noting the implications of the 
uses of the word "science" as part of the title of the new area. The article 
from which the following extract has been taken included a review of the 
first two volumes of the Administrative Science Quarterly which has been 
omitted in the following material. 

One of the most striking intellectual movements of the past generation 
was the development of specialized fields of study and competence in the 
applied social sciences. Public administration, business administration, and 
social work had developed professional schools by the 1930s, to which 
now must be added new varieties such as hospital administration and hotel 
administration. Industrial relations emerged as an independent discipline 

* From Administrative Science Quarterly, III (1958), 1-22. Reprinted by permission 
of the Administrative Science Quarterly. 


Evidences for an Administrative Science 1 1 9 

somewhere around 1930- 1940. International relations developed institutes 
of its own, and even economic development seems well on the way to be- 
ing an independent field of study. These "applied fields" have usually 
grown out of social science "departments" and at first are usually at- 
tached to them. Thus business administration and industrial relations tend 
to grow up within economics, public administration and international rela- 
tions within political science, social work and family relations within 
sociology, and so on. As soon as an applied field begins to separate itself 
from the parent "pure" discipline, however, it finds that it has to draw on 
all the social sciences for its theoretical substructures, its research methods, 
and its field of study. Thus industrial relations now has little more than 
a traditional attachment to economics; it draws its theoretical principles 
and its research methods from sociology, social psychology, and many 
other sources. The very process of developing an applied discipline com- 
pels the practitioner to search for his "pure" base in many different fields, 
both in theory and in research methods. 

One of the applied social science fields which has developed with great 
fertility is that which deals with the structure, functions, and behavior of 
organizations. Administrative science would seem to be a perfectly good 
name for this field; parts of it, however, go under other names— manage- 
ment science, public administration, business administration, operations re- 
search, organization theory, to name merely the most obvious. Two newly 
established journals— Management Science and Operations Research— oc- 
cupy at least part of the general field. The various journals put out by 
business schools cover a more specialized segment. One might even include 
the various activities which originate in and around schools of architecture 
under the general head of "planning." In addition to this outburst of fairly 
recent interest and activity there is an older movement, going back to 
Taylor and "scientific management" and to Gulick and Urwick and "pub- 
lic administration," which is so well established that by this time it is 
regarded as the "tradition" against which some of the newer movements 
are in revolt! 

At the outset we might frown on two possible discussions of a purely 
semantic nature which would not be particularly fruitful; the first is the 
question of what is the "right" name for this general area of study, the 
answer to this question being that any name which is not misunderstood is 
"right." Operations research covers those parts of the decision-making 
process in organizations which are most susceptible to mathematical treat- 
ment, where the elements in the problem are well adapted to measurement, 
and where the problem of the criterion of success in the operation either 
presents no difficulties or can be solved for the purposes of the project in 
hand by some reasonable but arbitrary assumptions. Management science, 

1 2 o Theoretical Aspects of Control 

as reflected in the journal and the society of that name, has a rather 
broader frame of reference and is interested in more general problems and 
theories than the "pure" operations researchers. It still retains a strong focus 
of interest in mathematical models of behavior, in "systems research," and 
in the quantification of organizational processes. Administrative science, 
as reflected in the Administrative Science Quarterly, is oriented more to- 
ward sociology, social psychology, and the social sciences "proper." It is 
less mathematical in the technical sense and is more concerned with the 
broad application of the theories and methods of the social sciences to 
problems of organization. These three approaches are complementary 
rather than competitive; each occupies a different part of the broad field 
of organizational studies, with overlapping boundaries. There is not the 
slightest point in arguing which is the "right" or even the "best" approach. 

The second useless controversy is whether any or all of these various 
studies deserve the holy name of "science." The question of whether a 
particular field of intellectual activity is "scientific" is seldom interesting. 
Knowledge is gained in all sorts of ways; methods which are appropriate 
in one area are not appropriate in others; and there is a large spectrum of 
more or less useful and knowledge-producing intellectual activities. It is 
unnecessary to label fields of study as more or less "scientific" and absurd 
to judge them by this label. Fields of study are more or less mathematical, 
more or less empirical, more or less experimental, more or less quantitative, 
more or less subjective, and so on, and in this many-dimensioned set each 
field must find that character which is most suited to its subject matter. 
The important question is not whether any field is "scientific" but whether 
its activities contribute usefully to the improvement of human knowledge 
and whether these activities might themselves be improved in various 
ways. If a purist for scientific method comes along to complain that ad- 
ministrative science is not "science," he can politely but firmly be shown 
the door,' and we can get on with the business of examining whether the 
content of administrative science is appropriate to its field of study and 
whether its methods usefully advance knowledge in that field. 

An economist can hardly fail to be impressed by the much higher level 
of generality in theoretical writing in the administrative science field by 
contrast with that in economics. The problem of the appropriate level of 
generality of theory is a difficult one; there is certainly need for work at 
many different levels. If a theory is to be fruitful in the sense that it points 
us toward conclusions which we would not have without it and toward 
empirical inquiry, which may either confirm it or force its reorganization, 
then it must find an "appropriate" level of generality. This must lie some- 
where below that of the large philosophical generalization which points 

Evidences for an Administrative Science 1 2 1 

to no conclusions and to no specific task of confirmation or refutation and 
somewhere above that of the ad hoc hypothesis which applies only to a 
particular case and has little or no significance outside a particular situa- 
tion. If a theory is to be fruitful, it must also develop an appropriate level 
of abstraction for its concepts and constructs. Thus the development of a 
new abstract concept, like, for instance, the Shannon concept of informa- 
tion, often leads to a very fruitful development of theory. The trick here 
is to be able to select from the vast complexities of real phenomena just 
those abstract properties which constitute the essence of the situation for 
the purposes of study. There are no rules, unfortunately, by which ap- 
propriate levels of generality and abstractness can be discovered auto- 
matically; the appropriateness of a theory can be known only by its fruits. 

The development of theory in administrative science is a formidable 
problem because of the inherent complexity of the field of study. Here 
we are dealing, not with simple mechanical systems, but with systems in- 
volving all the intricacies of human personality, images, and communica- 
tions. Any theoretical systems that we can construct will fall far short of 
the inherent complexity of the reality. An organization includes within its 
framework all systems levels. 1 Purely mechanical systems, like the theory 
of the firm in economics, are by no means useless but will take us only 
a very small part of the way. If organization theory has gone beyond this 
level, it is because of the recognition of cybernetic or control systems as 
an essential part of organization and of information as an abstract quantity 
linking the "roles" which form the pattern of the organization chart, 
formal or informal. 

Administrative science particularly needs good studies of exceptional 
individuals. Economics can get away with pretending that all men are 
much alike, for in the market place perhaps they are. As we move toward 
the study of organizations, and especially as we study the genesis and 
growth of organizations, the role of the exceptional individual, the role 
creator, the founder of religions, states, societies, and corporations, be- 
comes more and more important. Biography is therefore an essential part 
of the raw material of administrative science. We still need to do much 
more thinking on how to integrate the knowledge gained from descriptive 
and historical case studies into the knowledge gained from empirical and 
quantitative research. This is, I suppose, one of the problems of theory- 
how to provide an abstract image which is constantly enriched by both 
forms and sources of information. It would be optimistic to suppose that 
this problem had been solved. 

1 See K. E. Boulding, "General Systems Theory," Management Science, 2 (April 
1956), 197, for a discussion of systems levels. 

1 2 2 Theoretical Aspects of Control 

I turn now to the last category, that of empirical and quantitative re- 
search. If any single characteristic may be said to distinguish this form of 
activity from the preceding category of descriptive and historical work, 
it is the creation of "data." In descriptive writing the author takes a com- 
plex situation and tries to analyze it, as it were, in the raw. In "research" 
the investigator interposes a fairly standardized process of data collection 
between the situation and the analysis, and what he analyzes is not the 
situation but the data. The data may take the form of answers to question- 
naires, or they may consist of formalized observations of behavior em- 
bodied in the investigator's notes. In either form it is usually capable of 
some forms of quantification and statistical analysis. The end product con- 
sists of numbers which summarize in some way the properties of the data 
(indexes, coefficients of regression or correlation, chi-squares, tests of 
significance, and so on) which are arrived at by standard processes which 
are presumably understood by both writer and reader. The numbers may 
also be arranged in tables, which likewise summarize, in more detailed 
fashion, certain broad properties of the data. 

There are many advantages of this method. It is consistent with the 
methods of physical and biological science, which also abstract from the 
complexities of the "situation" a body of "data" which are then analyzed. 
When the situation is so complex as to defy analysis, it is a useful trick to 
substitute for it a body of data which can be analyzed. The principal 
danger of the method is that the investigator may be so absorbed in an- 
alyzing his data that he forgets entirely the situation out of which the 
data were abstracted. In the natural sciences this problem is not perhaps so 
serious as it is in the social sciences, for the world of nature is less complex 
than the investigator, and his abstractions, therefore, do less violence to 
the reality. In the social sciences, however, the investigator is dealing with 
situations of the same (or even of a higher) order of complexity than him- 
self. In this case the problem of the critique (not merely of the analysis) of 
the data is of great importance. Social scientists are much too prone to 
concentrate on the problems of the analysis of their data rather than on the 
relation between the data and the situation. This is why I suspect that the 
good social scientist should always be paired with a humanist, at least in- 
side his own head, who will constantly be looking at the situation as well 
as at the data and continually modifying the data-collection process in the 
light of increasing knowledge of the situation. 

The intellectual disease of analyzing data to the exclusion of the situa- 
tion may be called data fixation. Its principal symptom is a certain ob- 
sessiveness with arithmetic— the feeling that once a number has been ar- 
rived at by a recognized statistical ritual something has been accomplished. 
The article that is sandwiched between tables and peppered with coeffi- 

Evidences for an Administrative Science 1 2 3 

cients of correlation and statistical tests of significance is highly suspect in 
this regard. There are too many spurious quantities in social science re- 
search. I must confess that I regard the invention of statistical pseudo- 
quantities like the coefficient of correlation as one of the minor intellectual 
disasters of our time; it has provided legions of students and investigators 
with opportunities to substitute arithmetic for thought on a grand scale. 
Arithmetic is so much easier than thought that the temptation to make 
the substitution is almost irresistible. When the arithmetic is performed by 
electronic calculators, the substitution is even more disastrous. I recall a 
remark of the great statistician and econometrician Henry Schultz that 
the great value of statistical computation (this was in the days of hand 
computers) was that it got the investigator thoroughly familiar with his 
data. Now, alas, only the I.B.M. machine gets thoroughly familiar with the 
data; the investigator does not. His data are served up to him in a variety 
of digested forms, and he surveys a product, which is as far removed from 
the data as the data from the situation. 

It is the sense, I think, of a concerted effort by specialists from many 
different fields to solve a problem of great practical relevance to human 
welfare which creates the excitement— a sense also of new tools being ap- 
plied to old problems and a certain air of hope and enthusiasm that a new 
step forward in man's understanding of himself and his society is being 
taken. Euphoria is, of course, no substitute for wisdom, the proof of the 
pudding is yet to come, and one wants to protect oneself against dis- 
appointment by suitable head shakings and a sound conservative melan- 
choly; but for all this there is a feeling of dawn in the air, and it is good 
to be alive in it. 

The importance of the problem can hardly be exaggerated: how can 
organizations be built, and by what principles shall they be conducted, 
which will serve to free and not to enslave the individual and which shall 
be protected against the gangrene of corruption which besets all human 
institutions? The question goes back at least to the Greeks, and all human 
history is man's failure to answer it. It is not too much to hope that the 
plodding bricklaying of the social sciences, working with the larger but 
more unstable self-knowledge of man that comes from the humanities, 
may in our day yield some better answers to this question than the past has 
ever afforded. This is a high purpose, and our age is somewhat ashamed 
and suspicious of high purposes, so let us not talk about it. But with a 
high purpose tucked somewhere into the background, a firm resolve to 
gather contributions from many different disciplines, a certain amount of 
methodological skepticism, and a desire to brew a tasty combination of 
humanistic and social-scientific approaches, I venture to predict for the 
Administrative Science Quarterly a long and useful life. 

! 24 Theoretical Aspects of Control 


Herbert A. Simon* 

A leading authority discusses the major factors involved in developing 
a meaningful theory of organization. 


Human organizations are systems of interdependent activity, encom- 
passing at least several primary groups and usually characterized, at the 
level of consciousness of participants, by a high degree of rational direc- 
tion of behavior toward ends that are objects of common acknowledgment 
and expectation. Typical examples of organizations are business firms, 
governmental administrative agencies, and voluntary associations like 
political clubs. 

In complex enterprises the definition of the unit is not unambiguous— a 
whole agency, a bureau, or even a section in a large department may be 
regarded as an organization. In such a nest of Chinese blocks the smallest 
multi-person units are the primary groups; the largest are institutions (e.g., 
"the economic system," "the state") and whole societies. We will restrict 
the term "organization" to systems that are larger than primary groups, 
smaller than institutions. Clearly, the lower boundary is sharper than the 

Complexity in any body of phenomena has generally led to the con- 
struction of specialized theories, each dealing with the phenomena at a 
particular "level." Levels are defined by specifying certain units as the 
objects of study and by stating the propositions of theory in terms of 
intra-unit behavior and inter-unit behavior. (Cf. the sequence of ele- 
mentary particle-atom-molecule in physics and .the sequence: gene-chro- 
mosome-nucleus-cell-tissue-organ-organism in biology.) 

Not every arbitrarily selected unit defines a suitable level for scientific 
study. The most important "unities" that make a level an appropriate one 
for theory construction and testing appear to be the following: 

(a) The units at the level in question should exhibit a high degree of 
internal cohesion relative to their dependence on each other. Under these 
circumstances we can discover generalizations about the internal proper- 
ties of the individual units as quasi-isolated systems (e.g., propositions 
about communications patterns among component primary groups of an 

* From The American Political Science Review, XL VI, 4 (1952), 1130-40. Re- 
printed by permission of the American Political Science Association. 

Comments on the Theory of Organizations i 2 5 

organization). We can also discover approximate generalizations about the 
relations between units as wholes (e.g., propositions about competition be- 
tween two organizations) . 

(b) The units should exhibit internal properties that are different (or 
depend on different mechanisms) from those that predominate in the in- 
ternal properties of sub-units at the next level below (e.g., the determi- 
nants of the volume of communication between members of two different 
primary groups in an organization should be distinguishable in important 
respects from the determinants of the volume of communication between 
members of a single primary group). 

These two tests are not intended as metaphysical assertions about 
"wholeness" or "emergent" properties, but simply as criteria determining 
whether, in fact, verifiable propositions can be constructed employing the 
units in question as approximations to the full complexity of nature. Even 
if at some stage in inquiry we should be able to reduce the propositions 
of theory at one level to those at the next lower level— as the theory of 
gases has been reduced to statistical mechanics— the former propositions 
would still retain their usefulness for purposes of application and economy 
of statement. Indeed, the value of both sets of propositions is enhanced by 
their translatability from the one to the other. 

Human organizations would seem to qualify to a high degree as suitable 
units defining a level of analysis of systems of human behavior. With re- 
spect to the first criterion stated above, the most superficial observation 
shows that the boundaries between organizations have real behaviorial sig- 
nificance, and that it is meaningful, in first approximation, to state proposi- 
tions about the relations between organizations regarded as wholes. (I 
trust that I have made clear that no notion of "group mind" is implied in 
this last statement.) 

With respect to the second criterion, I believe that enough is known 
about the psychological mechanisms that are primarily responsible for 
cohesion and interdependence in the primary group to show that these 
mechanisms cannot easily account for the corresponding phenomena in 
the larger organized aggregates; and that there are important organiza- 
tional phenomena that do not have exact counterparts at the primary 
group level. A number of examples of these mechanisms and phenomena, 
which are central to organizations but absent from or of lesser importance 
to primary groups, will be given in the next section. 

But why speak of a level of organization theory? Do we not need as 
many levels as there are structural layers between primary groups and 
institutions? I think not, because I do not believe that these various levels 
are distinguishable to an important extent in terms of the second criterion 
suggested above— i.e., there are no important new mechanisms to be dis- 
covered at these successive levels. The propositions of organization theory 

1 2 6 Theoretical Aspects of Control 

can probably be stated with systematic ambiguity so as to refer indiffer- 
ently to the relations of divisions within a bureau or the relations of 
bureaus within a department. As small differences in degree begin to ap- 
proach qualitative significance at the upper end of the scale, we have 
probably already reached the level of institutional theory. In the future, 
of course, the results of research may force us to revise this assumption 
and to introduce additional levels of theory. 


The study of organizations has hardly progressed to the point where a 
definitive list can be constructed of the major areas for research. The fol- 
lowing list was arrived at primarily by considering which characteristics 
of organization—particularly those distinctive ones that identify the level 
of organization theory— require dissection and explanation. I have not tried 
to construct watertight categories, and it will become evident that several 
of the items represent different ways of looking at the same problem. Until 
we know what frames of reference are going to be the most useful for 
organization theory, it will surely be desirable to retain alternative frame- 
works, and to take considerable pains to develop means for translating 
from one framework to another. 

z. The process of decision-ma kin gdn organization. A language for the 
description of decision-making processes appears to offer considerable 
promise as a framework for the study of organizations. The central notion 
is that a decision can be regarded as a conclusion drawn (though not in 
any strict logical sense) from premises; and that influence is exercised by 
transmitting decisions, which are then taken as premises for subsequent 

When the problem of influence is stated in these terms, our attention is 
called to some features that are not prominent in other formulations. We 
see, for example, that the process may depend not only upon interpersonal 
relations between influencer and influencee, but also upon the structure 
and accepted rules of transformation of the language employed by them. 
One can begin investigation here by posing such questions as how in- 
fluence is transmitted in an organization between professional groups that 
employ different problem-solving technologies, e.g., accountants and engi- 
neers. Work on organization theory utilizing this framework could prob- 
ably soon be related, in a mutually beneficial way, to research on the 
sociology of knowledge and on the psychology of the problem-solving 

2. The phenomena of power in organizations. A characteristic feature 
of the mutual influence of organization members upon one another is that 
this influence exhibits striking asymmetries-as, for example, in the 

Comments on the Theory of Organizations 1 2 7 

superior-subordinate relationship. These asymmetries appear to be what 
we have chiefly in mind in using such terms as "power" and "authority." 
The following are a number of important research tasks in this area. 

a. A fully operational definition of power and methods for observing 
and measuring power relationships is not yet at hand, but would seem 
fundamental to the description of organizational behavior. 

b. More needs to be learned about the motivational basis of power in 
organizations, including the roles of sanctions, identifications, and atti- 
tudes of legitimacy in the acceptance of authority. Progress has been made 
in the study of the analogous phenomena in primary groups (e.g., work 
on leadership and on group morale), but it is not obvious that the mechan- 
ics of influence within the primary group tell all, or even most, of the 
story of influence processes in larger organized aggregates. 

c. In elaboration of the last point, the distinction between the "formal" 
and the "informal" in organizations appears to lie, in part, in differences 
between the psychological bases of cohesion that are involved. When we 
refer to power as formal, what we appear to mean is that internalized 
attitudes toward legitimate authority provide the motivation for accept- 
ance of the relationship. While feelings about legitimacy undoubtedly 
play a role in primary group relationships, I would conjecture that they 
take on additional importance when they serve as a substitute for the 
immediate experience of approval and disapproval in face-to-face relation- 

d. Another mechanism that is important in the transmission of influence 
in organizations is the interlocking of primary groups through the dual 
membership of supervisory employees. In general, each supervisory em- 
ployee is a member both of a group in which he is formal leader and of 
another in which his immediate superior is formal leader. The principal 
research problems here are to determine the behavior patterns that are 
adopted by executives in these "cross-pressure" situations; and, if there 
are several such patterns, to find what determines which one will be 
adopted. The same questions need to be answered with respect to the 
"staff" man who, because he is attached to a "line" unit, also has potential 
or actual membership in two primary groups. It remains to be seen 
whether cross-pressures produce the same behavior in these organizational 
situations as in the other social situations where they have been studied. 

5. Rational and non-rational aspects of behavior in organization. Or- 
ganizations are the least "natural," most rationally contrived units of 
human association. But paradoxically, the theory of an organization whose 
members are "perfectly rational" human beings (capable of unlimited 
adaptation) is very nearly a perfectly vacuous theory. It is only because 
individual human beings are limited in knowledge, foresight, skill, and 
time that organizations are useful instruments for the achievement of 

1 2 8 Theoretical Aspects of Control 

human purpose; and only because organized groups of human beings are 
limited in ability to agree on goals, to communicate, and to cooperate that 
organizing becomes for them a "problem." 

Organization theory is centrally concerned with identifying and study- 
ing those limits to the achievement of goals that are, in fact, limits on the 
flexibility and adaptability of the goal-striving individuals and groups of 
individuals themselves. The entrepreneur of economic theory is limited 
only by constraints that are external to himself and his organization— the 
technology— and by the goal-striving of individuals whose interests are 
not identical with his. Administrative man is limited also by constraints 
that are part of his own psychological makeup— limited by the number 
of persons with whom he can communicate, the amount of information he 
can acquire and retain, and so forth. The fact that these limits are not 
physiological and fixed, but are instead largely determined by social and 
even organizational forces, creates problems of theory construction of 
great subtlety; and the fact that the possibilities of modifying and relaxing 
these limits may themselves become objects of rational calculation com- 
pounds the difficulties. 

In this general area of research, promising suggestions as to the direction 
in which we might move are contained in oligopoly theory and game 
theory (formulation of the "outwitting" problem), and in sociological 
speculations about the self -confirming prophecy. I would single out the 
following areas for special attention: 

a. Identification of the limits of rationality. We need a more complete 
and systematic taxonomy of the constraints, internal to the system of 
social action, that serve as limits to the attainment of goals. This would 
lead to empirical research on the questions: (i) under what circumstances 
particular constraints do and do not operate, including inter-cultural uni- 
formities and differences, and (ii) under what circumstances the modifica- 
tion of removal of particular constraints becomes an object of rational 

b. Theory of organizational innovation and change. Plans are regarded 
as "utopian" when their implementation would require changes in internal 
constraints that are thought to be unchangeable. Essentially, Utopian plans 
are rejected because "you can't change human nature" in those respects 
that would be essential to achievement of the plan. Research is needed 
as to the criteria that are applied by human beings in planning situations 
to determine which of the behavior variates they will regard as variable 
(i.e., subject to rational determination), and which as fixed (i.e., con- 
straints on goal attainment) . 

c. Reification of groups. The limit of human understanding in the pres- 
ence of complex social structures leads human beings to construct simpli- 
fied maps (i.e., theories or models) of the social system in which they are 
acting, and to behave as though the maps were the reality. To the extent 

Comments on the Theory of Organizations 129 

that such maps are held in common, they must be counted among the in- 
ternal constraints on rational adaptation. What we have just said applies, 
of course, to all systems of classification which, by determining when 
situations are "similar" and when "different," provide the individual with 
the social definition of the situation. 

My earlier comments about the relation of "formal" organization to at- 
titudes of legitimacy can be generalized in terms of this notion of social 
classification. The process of organizing involves, among other things, 
securing acceptance by the organization members of a common model 
that defines the situation for them, and provides them with roles and ex- 
pectations of the roles of others, and with commonly accepted classifica- 
tory schemes. Attitudes of legitimacy probably provide a principal moti- 
vational base for the organizing process. 

What is needed here is study of the factors that determine how an 
organization will be perceived by the persons in it, how the mode of per- 
ception affects behavior, and what the effects are of a greater or lesser de- 
gree of sharing of such perceptions. 

4. The organizational environment and the social environment. Mem- 
bers of an organization generally come to it already equipped with the 
mores of the society in which it operates. To what extent can and do 
organizations develop and inculcate mores that are distinct from the mores 
of the society? To what extent are there in a society generalized mores 
about behavior in organizations that provide the basis for the operation 
of the individual organizations in the society (e.g., generalized mores 
about superior-subordinate roles)? 

Organization theory has been largely culture-bound through failure to 
attack this problem. The theory of bureaucracy as developed by Max 
Weber and his followers represents the furthest progress in dealing with 
it. The historical data appealed to by the Weberians need supplementation 
by analysis of contemporary societies, advanced and primitive. A com- 
parison of intracultural uniformity and variation in organization patterns 
with inter-cultural uniformity and variation would provide the evidence 
we need to determine to what extent the cooperative patterns in organiza- 
tions are independent of the mores of cooperation of the society. 

$. Stability and change in organizations. Any theory of the movement 
of a system of organizational behavior through time must take account of 
the apparent stability exhibited by organizations. From every evidence, 
this stability must be an extremely complex phenomenon. It may rest in 
part on the kinds of bonds, which we might refer to as non-rational, that 
have been observed in the primary group; it may depend in part on the 
rational calculations of members that their interests are served by the 
organization. It is because the role of these, and possibly other, bases of 
stability needs to be explored that I offer the following suggestions: 

! . Theoretical Aspects of Control 

a. It is possible that systems in which the "non-rational" type of stabiliz- 
ing mechanism predominates will behave in a qualitatively different fashion 
from those in which the "rational" type of stabilizing mechanism prevails. 
If, by construction of models embodying the two types of mechanisms, a 
qualitative difference could be deduced, the way would be open to em- 
pirical assessment of the importance of the two mechanisms. 

b. The work that has been done to date on the theory of the "rational" 
mechanism would suggest that stability in this case depends on certain 
relations between the aspiration levels of members and their achievement 
levels. If so, we can draw on the psychological research that has already 
been done on these latter phenomena to design experiments and field 
studies that would test whether this is, indeed, one of the mechanisms 
involved in stability. 

c. We may borrow the economists' term "entrepreneur" to refer to an 
individual who specializes as a broker in finding mutually acceptable terms 
on which a group of persons can be induced to associate, or to continue 
association, in an organization. We need research to determine what the 
role is of entrepreneurship, so defined, in the process of organizational 
activity. I conjecture that there are some close relationships both with the 
"middleman" notion, introduced in topic 2 d, and with the kind of sta- 
bility mechanisms discussed in 6 b. Study is also needed of whether the 
uniqueness or non-uniqueness of the acceptable terms of association is an 
important determinant of the amount of authority that can be exercised 
over organization members. This relationship has been exhibited in some 
formal models, but it needs empirical verification. 

d. The two topics just discussed get very close to the heart of the proc- 
esses of bargaining and the formation of coalitions, insofar as these 
processes involve rational calculation of advantage. The formal apparatus 
of game theory appears to provide an appropriate language of theory 
formulation, and, on the empirical side, some of the problems could prob- 
ably be examined by means of relatively small-scale laboratory experi- 

e. Another aspect of survival and stability is the question of how or- 
ganizations adapt themselves to uncertainty and incomplete information. 
In the past two decades this has been a favorite topic of economists, but 
only in the last five years has there been much attention to the two aspects 
of greatest importance to organization theory: (i) reduction in the impact 
of uncertain events by retention of "flexibility" and (ii) the role of a 
stable social environment as a means of providing predictability to the 
individuals who are a part of it. 

Under the first heading, research is needed as to the implications of 
particular ways of organizing behavior for the adaptability of the organi- 
zation under changing, unpredictable circumstances. Under the second 

Comments on the Theory of Organizations 1 3 1 

heading, research is needed as to the existence and nature of mechanisms 
in social organizations that are analogous (in the sense of performing the 
same function) to the homeostatic mechanisms of organisms. Whether or- 
ganizations are adaptive and possessed of homeostatic mechanisms is an 
empirical question, but one which, in all probability, can be answered in 
the affirmative. But the important theoretical issue is the nature of the 
mechanisms— a question that is not solved by reference to the organistic 
analogy. Moreover, while primary groups and social institutions may also 
exhibit homeostasis and adaptivity, there is no reason to believe that the 
mechanisms involved are the same ones that produce these phenomena 
in organizations. Functional equivalence does not imply structural equiva- 

6. Specialization and the division of work. The division of work and 
the design of the organizational communications system have in the past 
been the central concerns of persons interested in organization theory for 
purposes of application. The question usually asked is: "How do we di- 
vide the work, and what channels of communication do we establish in 
order to operate efficiently?" For purposes of research, the question is 
more properly stated: "What are the consequences for organizational 
activity of dividing the work one way rather than another, or employing 
one set of communications channels rather than another? " 

The last half of the question (communications) is best answered in 
terms of the frames of reference of topics 1 and 2. The subject of the 
division of work requires further comment. We are considering, of course, 
not only the question of specialization of the individual organization 
member, but also the allocation of tasks to whole organization units— in 
fact, it is the question of specialization among the larger aggregates rather 
than specialization within the primary group that is the proper concern 
of organization theory. We are equally concerned with "vertical" spe- 
cialization— i.e., allocation of decision-making functions to various status 
and authority levels in an organization— and with "horizontal" specializa- 
tion— i.e., fixing the jurisdictional boundaries of co-ordinate organizational 

a. Current theories of specialization in organization (excluding the 
"human relations" approach to the primary group) are largely derived, 
via the scientific management movement, from Adam Smithian notions 
that specialization is a means to efficiency and hence to effective compe- 
tition. There has been little examination of the alternative Durkheimian 
idea that specialization is a means of protection from competition. The 
research problem suggested by the contrast is to examine in what respects 
specialization (and what kinds of specialization) increases organizational 
stability; in what respects it jeopardizes stability; and to what extent 
these considerations enter into decisions about specialization. The problem 

! 3 2 Theoretical Aspects of Control 

is also related to 5 e in that certain forms of specialization may make an 
organization less dependent on what other organizations do, and hence 
may provide a means for dealing with uncertainty. 

b. The consequences of specialization depend on the constraints dis- 
cussed in topic 3. It is an important question as to how far specialization 
is determined by constraints external to the organization— the technology 
of its activities— and how far it is determined by internal constraints— the 
psychological and sociological limitations upon rational adaptation. (The 
situation is even a bit more complicated because the technology— in the 
sense of the physical, chemical, biological, etc., processes involved in the 
organization's activity— is not independent of the state of technological 
knowledge, and the latter may, in turn, be interdependently related to the 
forms of social specialization that prevail.) In almost every city, the fire 
department is a recognized organization unit, and in almost every steel 
mill, the blast furnace department. Here are examples of specialization that 
appear to be dictated by the technology— the units are "natural" in this 
sense. On the other side we find units that are "natural" in the sense of 
being specialized to handle socially-defined purposes, which, in turn, de- 
pend on the processes of reification discussed in 3 c (e.g., the Children's 
Bureau). Research into the theory of specialization making use of the 
framework suggested in topic 3 is needed to clarify these issues, and to 
formulate and test propositions about the consequences of specialization. 

c. The relationship between specialization and the internal constraints 
on rational adaptation is two-way. The division of work may be de- 
termined, partly or wholly, by such constraints; it will in turn create 
constraints. That is, the form of specialization will be a major determinant 
of the frames of reference, skills and knowledge, identifications and foci 
of attention of organization members. Probably this is the most promising 
viewpoint from which to tackle the non-rational aspects of formation of 
group identifications (or "interests" in the political sense) and the effects 
of such identifications upon inter-group processes (cf. 5 d on the "ra- 
tional" aspects). 

d. Problems of vertical specialization are closely related to topics 1 and 
2. In applied organization theory, the questions are usually stated in terms 
of "centralization" and "decentralization." 

This list of research areas illustrates, I think, that the phenomena of or- 
ganization constitute an important level of theory— a level that is encom- 
passed neither by the usual conceptualizations of small-group processes 
nor by those of the more macroscopic analyses of cultures and institutions. 

The characteristics of this level that give it its particular "flavor" are 
the following: (a) its focus is on relations among interlocking or non- 
interlocking primary groups rather than on relations within primary 

Theory of Organizational Behavior 1 3 3 

groups; (b) it is largely concerned with situations where znjoeckrationalitat 
plays a large role relative to wertrationalitat (as compared with the study 
either of small groups or of cultures); (c) in these situations the scheme 
of social interaction becomes itself partly a resultant of the rational con- 
triving of means and the conscious construction and acting out of "artifi- 
cial" roles; and (d) explanation of phenomena at this level requires the 
closest attention to the fluid boundaries of rational adaptation, including 
the important boundaries imposed by group frames of reference, percep- 
tual frameworks, and symbolic techniques. In contrast to these characteris- 
tics, the level of primary group theory must pay much more attention to 
the personal values that are emergent from the process of group inter- 
action itself, the acculturalization of individuals to the group, and the par- 
ticular forms of cohesion that arise out of face-to-face interaction and in- 
dividual sensitivity to group approval. 

It would be wrong, of course, to insist that none of the primary group 
phenomena are relevant to inter-group relations, or vice versa. Neverthe- 
less, the important work that has been done on small groups in the past 
generation— much of it involving the observation of groups that were part 
of larger organizational structures— has contributed very modestly to the 
solution of the problems of organization theory. 



Robert V. Presthus* 

Professor Robert V. Presthus explores theoretical formulations from 
sociology and psychology in an attempt to set down a general theory of 
organizational behavior. The article is of particular interest because of its 
insight into the factors entering into the structuring and functioning of an 
organization such as a large business unit. 

During the recent past the analysis of organization has shifted from a 
preoccupation with structured rationality to an emphasis upon individual 
behavior. Much of this emphasis has been sociological, that is, it is con- 
cerned mainly with small groups and with the ways in which such groups 
shape the alignment and use of power in the organization. The following 

* From Administrative Science Quarterly, III (1958), 48-72. Reprinted by permis- 
sion of the Administrative Science Quarterly. 

^34 Theoretical Aspects of Control 

analysis attempts to add another dimension to this main drift by pulling 
together the insights of several social sciences into a general theory of or- 
ganizational behavior. Such efforts seem required, however crude and 
abstract they may be at this early stage in the development of administra- 
tive science. 

In line with Merton's plea for more attention to the interplay between 
bureaucratic structure and personality, 1 some psychological formulations 
are brought to bear upon two major variables, the total organizational 
situation and the individual. Such a framework seems well suited to the 
complexity of organizational behavior, which is the product of interaction 
among the whole culture, a given organization, and an individual per- 
sonality which itself is the result of the genetic composition and unique 
experience of any given individual. In this context an organization may be 
viewed as a miniature society in which traditional social controls over 
the individual appear in sharp focus. The organization draws upon the 
accumulated learning and experience of the individual, who brings to it 
certain socially inculcated attitudes that encourage a satisfactory accom- 
modation to the organization's major values and expectations. Obviously 
not all individuals achieve this kind of accommodation, but the vast 
majority do so at varying levels of identification and self-realization. 

Without denying the influence of informal, small-group liaisons, we as- 
sume that individuals have several reference points other than their im- 
mediate work group, including the organization as a whole. This concept 
of differentiated reference foci is suggested by the dichotomy between 
"cosmopolitans" and "locals," between those whose loyalty is bound up 
with their own organization (locals) and those whose referential context 
is profession- wide and national (cosmopolitans). 2 Here we are concerned 
mainly with "locals" who tend to accommodate successfully. While our 
theory necessarily includes an "ambivalent" type who tends to reconcile 
inapposite adjustment by resignation, aggression, and withdrawal, this in- 
quiry is directed toward those "upward-mobiles" and "indifferents" to 
whom the organizational bargain is either satisfactory or at least insuffi- 
ciently unsatisfactory to provoke disengagement. The question thus be- 
comes, What aspects of the dynamic interplay between the total organiza- 
tional situation and the individual encourage these different kinds of 
accommodation? Part of the answer seems to lie in the individual's percep- 
tion of the organization as a social instrument and in the ways that the 
organization engages the deep-seated attitudes of the individual toward 

1 Robert K. Merton, "Bureaucratic Structure and Personality," Social Forces, 17 
(1940), 560-68. 

2 Cf. Alvin W. Gouldner, "Cosmopolitans and Locals: Toward an Analysis of 
Latent Social Roles," Administrative Science Quarterly, 2 (Dec. 1957-March 1958), 
281-306, 440-80. 

Theory of Organizational Behavior ! 5 r 

In this general framework, organization is defined as a system of struc- 
tured interpersonal relations, that is, individuals are differentiated in terms 
of authority, status, and role with the result that personal interaction is 
prescribed or "structured." Anticipated reactions tend to occur, while 
ambiguity and spontaneity are decreased. It is hypothesized that the re- 
sultant psychological field has exceptional influence upon learning and 
accommodation to the organization. 3 A related hypothesis is that behavior 
will tend to be more predictable in complex, structured organizations than 
in so-called voluntary associations. These assumptions reflect Harry Stack 
Sullivan's interpersonal theory of psychiatry and particularly his view 
that "the human organism is so extraordinarily adaptive that not only 
could the most fantastic social rules and regulations be lived up to, if they 
were properly inculcated in the young, but they would seem very natural 
and proper ways of life." 4 It would seem that the rational character and 
demands of the typical big organization will surely appear less than "fan- 
tastic" to its members. 

According to Sullivan, personality is the result of a "self-system" 
worked out through successful (anxiety-reducing) accommodations to 
the wishes of successive authority figures, such as parents, teachers, super- 
visors, and so on. The theory of anxiety is central, since anxiety is the 
principal medium by which the individual is exposed to the values of those 
in authority. Sullivan's conclusions as to the significance of the anxiety- 
conformity-approval syndrome can be summed up as follows: "I believe 
it fairly safe to say that anybody and everybody devotes much of his life- 
time and a great deal of his energy ... to avoiding more anxiety than he 
already has, and if possible, to getting rid of some of this anxiety." 5 

It is assumed here that anxiety is probably the most critical variable in 
organizational behavior, when such behavior is defined as an interpersonal 
process occurring in a highly structured environment. Such behavior is 
always associated with individual reactions to authority, which in turn 
are mediated by anxiety and the structure of the immediate interpersonal 
situation. It is important to add that anxiety occurs along a continuum, 

3 By "organization" I mean the ideal-typical bureaucratic model, characterized by 
large size, hierarchy, specialization, centralized formal power, and an orientation 
toward written rules and tradition as the main guides to behavior. Examples include 
a government bureau, a large university, an industrial concern with something over 
one thousand employees, an army command at the regimental level or above. Mem- 
bers of these organizations are usually selected on the basis of technical skill; typically, 
they view their vocation as a career, with clearly defined avenues. This model is 
highly idealized, of course, but perhaps no more so than the economist's "law" of 
supply and demand, which has proved quite useful for analytical purposes. 

4 H. S. Sullivan, The Interpersonal Theory of Psychiatry (New York, 1953), p. 6. 

5 H. S. Sullivan, "Tensions, Interpersonal and International," in H. Cantril, ed., 
Tensions That Cause Wars (Urbana, 111., 1950), p. 95. 

j -. 6 Theoretical Aspects of Control 

ranging from extreme, disorganizing fear to the natural uneasiness felt 
by most people in a strange situation. For our purposes, anxiety is de- 
fined as an unpleasant tension that, in Sullivan's terms, guides the develop- 
ment of the self -system, is present in some measure in all interpersonal 
relations, and is the main influence determining how such relations de- 
velop. This degree of tension may be called "adaptive anxiety" because in 
most cases it facilitates personal accommodation. In terms of organizational 
needs, such anxiety and the "security operations" that seek to overcome 
it are usually functional. 

This is not to say that anxiety reduction is the only motive for ac- 
commodation. Individuals seek opportunities for joy, love, self-realiza- 
tion, and power, which are not necessarily tied to anxiety, although they 
may be. Certainly, because power in our society can usually be equated 
with the control of organized resources, organizations provide unusual 
opportunities to satisfy this drive. But for the majority such expansive 
states as deep emotional satisfaction, love, and self-realization are not 
usually obtainable within the organization, which instead tends to stifle 
the spontaneous, idiosyncratic satisfactions that individuals seek. The 
"professional mask," the pleasant detachment, the rivalry, and the anxiety 
that characterize interpersonal relations in big organizations are germane. 
For most of us, impersonality, limited discretion, built-in power inequi- 
ties, and the fact that work often becomes a means of buying more mean- 
ingful ofT-the-job satisfactions suggest that less positive motives such as 
anxiety reduction warrant closer analysis. 

Something may now be said about the relation of anxiety to learning. 
Anxiety is apparently a kind of free-floating dread that affects most inter- 
personal relations to some degree. Unlike fear which has objective refer- 
ents, anxiety is often vague. Moreover, whereas fear usually relates to 
physical injury, anxiety relates to threats against personal esteem. But 
anxiety also has a positive role: it facilitates learning by sharpening both 
motivation and perception. As Sullivan concludes, "The first of all learn- 
ing is called out to avoid recurrence of the extremely unpleasant tension 
of anxiety, which is, and always continues to be, the very antithesis of 
everything good and desirable ... the child soon learns to discriminate 
increasing from decreasing anxiety and to alter activity in the direction 
of the latter. The child learns to chart a course by the anxiety gradient." 6 
Broadly, then, this analysis is grounded in the environmental school of 
psychology, which, without denying the biological foundation of human 
behavior, believes that cultural values play a major role in determining 
man's character. While this school is a minority one in psychology, its 

*lbid.; also see John Dollard and N. E. Miller, Personality and Psychotherapy 
(New York, 1950), p. 190; O. H. Mowrer, "Anxiety Reduction and Learning," 
Journal of Experimental Psychology, 27 (1940), 497-516. 

Theory of Organizational Behavior 1 3 7 

formulations seem to me to be most useful in analyzing organizational 
behavior. Believing that social institutions shape behavior, this school turns 
to anthropology and sociology as well as biology to understand man. The 
theory and research of cultural anthropology reinforce the view that 
environmental conditions largely shape individual behavior and personal- 
ity, since biological impulses such as pugnacity and competitiveness take 
quite different forms in different societies. The attending emphasis upon 
anxiety follows, since it becomes the primary mechanism for exposing the 
individual to cultural pressures. In the context of organization, it is sig- 
nificant that both Fromm and Thompson agree that the so-called "mar- 
keting character," who can be equated with the successful "other-di- 
rected" organization man, will "often tend to automaton conformity." 7 
Finally, since individual reactions to organizational stimuli are the result 
of learning, the concepts of perception and reinforcement are enlisted 
to help explain behavior in a structured environment. 

These several formulations underlie the present assumption that the 
patent status and power apparatus of organizations sharpens anxiety and 
thus increases the probability that behavior will reflect organizational 
premises. Complex organizations have an exceptional influence upon in- 
dividual behavior because they are organized systems of expectation. 
Their status and authority symbols function as patterns of manifest stimuli 
that reinforce the human tendency to honor majority values. The prob- 
ability of compliance is increased by the fact that organizational behavior 
is group behavior of an exceptionally structured kind. 

All human behavior occurs within some normative framework, con- 
sciously articulated or (more frequently) tacitly assumed. In big organiza- 
tions there is a fairly consistent hierarchy of values, culminating in a final 
ideal, the "good of the organization." Among the advantages of such a 
criterion is its flexibility and ambiguity— it can be manipulated to meet 
most exigencies. Because the organization must always compete for popu- 
lar approval, consumer loyalty, and legislative protection and because its 
power must be constantly nourished, its major values become important 
tactical instruments. They are personified by the leaders of the organiza- 
tion and explicit in its traditions, and they provide behavioral cues for its 
rank and file. The assumed best interest of the organization thus provides 
a standard for determining policy, evaluating individual performance, 
defining loyalty, and rationalizing injustice if injustice becomes necessary. 
Obviously this standard may be misapplied in any given instance, but this 

7 C. Thompson, Psychoanalysis: Evolution and Development (New York, 1950), 
p. 208; also, for an excellent, methodologically rigorous work which interprets be- 
havior in anxiety-reduction terms, see Timothy Leary, Interpersonal Diagnosis of 
Personality: A Functional Theory and Methodology for Personality Evaluation (New 
York, 1957). 

! ^ g Theoretical Aspects of Control 

possibility is not significant here because the decision makers always try 
to apply it rationally. 

Imperatives such as these are reflected in the psychology of the organiza- 
tion. They result in a conscious effort to increase its predictability and 
internal discipline. The organization tends to become a routine of skill, 
energy, and opinion. The structured interpersonal relations with which 
this analysis is concerned are part of this rational climate. They increase 
the probability that individual behavior will reflect organizational neces- 
sity. The individual is conditioned to accept the legitimacy of obedience, 
for example, by the very fact that he has been hired to do a specific job 
with explicit obligation, by the provision of rules and regulations that 
limit his discretion, and by the definition of his place in a hierarchy of au- 
thoritative relationships. But such situational factors become most mean- 
ingful in terms of the psychological impact they have upon members of 
the organization. The resultant attitudes, it must be said, obviously help 
the organization achieve its objectives, but they also have certain unantici- 
pated consequences that are dysfunctional in terms of such goals and of 
personal adjustment. 8 

We turn first to learning theory, because individual accommodation 
to the organization is essentially a matter of learning. Learning may be 
defined as a modification of behavior resulting from repeated exposures 
to a certain kind of stimuli. Learning proceeds according to a stimulus- 
response mechanism; its effectiveness depends upon various factors, in- 
cluding the number and strength of existing habits, perception, and the 
strength of the drive evoked by the stimulus. Perception is the process 
of becoming acquainted with the environment. Its motives include anxiety 
and what seems to be an instinctive tendency to use our sense organs 
functionally. 9 Random observation suggests that we appraise new social 
situations in order to orient ourselves, to decide what role is required. Our 
perception of a situation defines our behavioral limits in the sense that its 
speed and accuracy determine the appropriateness of the role we choose. 

Obviously differences in intelligence, emotional maturity, and motiva- 
tion influence perception and behavior. Some individuals have a limited 
social sensitivity, that is, their reactions are inappropriate to the "normal" 
expectations of a given situation. Such behaviors reflect, in part, inade- 
quate or distorted perception. Among higher animals, whose perceptual 
organization is acquired (learned), the sensitivity toward, range of, and 
discrimination among stimuli are greater than in lower animals, who de- 
pend largely upon inherent perceptual facility. Man is highly susceptible 
to learning because he is more aware of stimuli and more selective than 

8 Merton, op. cit.; C. Argyris, "The Individual and Organization: Some Problems of 
Mutual Adjustment," Administrative Science Quarterly, i (June 1957), 1-24. 

9 S. S. Stevens, Handbook of Experimental Psychology (New York, 1951), pp. 

Theory of Organizational Behavior 1 39 

other animals. It is also clear that perception is bound up with environ- 
ment, since the latter provides the potential stimulus field. This leads to a 
basic assumption: In the structured milieu of a big organization, we can 
assume that both perception and conditioning are facilitated by the mani- 
fest, authoritative nature of the stimuli. 

Reinforcement is also vital to learning, since it makes possible condition- 
ing through rewards and punishments. We know that responses which are 
followed by reinforcement will be learned; they will result in changes 
in the individual's behavior or response patterns. Individuals develop cer- 
tain tensions reflecting needs for food, water, sex, recognition, power, 
security, and so on. The behaviors which satisfy these needs are reinforced 
because they reduce the tension generated by the need. The reduction of 
tension is thus an unusually powerful reinforcement. Learning is also 
mediated by attitudes, that is, dispositions to act in a certain way. We 
learn things that agree with our preconceptions much easier and retain 
them longer than those that seem alien. Through this process of selective 
perception a steady reinforcement of accepted values occurs. The rein- 
forcement that accompanies the reduction of other tensions also operates 
with regard to the anxiety reduction achieved by deference to authority. 

While there is disagreement as to the applicability of their findings to 
human conduct, Pavlov and Skinner have shown how problem-solving in 
dogs and rats can be conditioned through the manipulation of stimuli and 
the use of rewards and punishments. 10 Pavlov's experiments demonstrated 
the conditioned response by adding a new stimulus, a bell sound, to the 
environment of dogs in a feeding situation. Normally, dogs salivate only 
when being fed or upon seeing food, but after the bell was rung many 
times just before the dogs were fed, Pavlov found that the sound itself 
caused salivation. The animal had become conditioned to the bell sound. 
Skinner's experiments with rats illustrate the importance of reinforcement 
in learning, using the reward principle. A hungry rat, placed in an empty 
box equipped with only a lever which releases food to him, will in time 
depress the lever. The consequent reduction of hunger increases the 
chances that the rat will repeat the action. This is a nice example of the 
need-reduction principle in behavior or learning. 

Reinforcement and motivation, building upon individual needs and the 
perceived means to meet them, lead to learning and habit formation. 
Through learning, the individual gradually selects from among several 
potential behaviors those which seem to have the best consequences in a 
given situation. The rationality of his choices will vary according to his 
intelligence, knowledge, information in a given situation, social percep- 
tivity, and so on. In a bureaucratic setting, the predictability of behavior 
and the probability that it will be functional in organizational terms are 

10 I. P. Pavlov, Conditioned Reflexes (Cambridge, Eng., 1927) ; B. F. Skinner, The 
Behavior of Organisms (New York, 1938). 

j 4Q Theoretical Aspects of Control 

greatly enhanced by the limited number of behavioral alternatives. Func- 
tional responses are reinforced by rewards. In bureaucratic occupations 
an obvious example is the granting of frequent yet small pay increases. We 
can assume that the stimuli which elicit desired reactions are manipulated 
in terms of organization needs. Such manipulation occurs at various levels 
of sophistication, but the harnessing of social science research to com- 
mercial objectives, advertising and sales psychology (e.g., the impact of 
repetition), opinion surveys, consumer research, the professionalization 
of charity fund drives, and so forth suggests the growth in the systematic 
molding of human behavior. 

Something must now be said about authority, which is among the main 
conditions of organized behavior. Authority is usually defined as the 
ability to elicit compliance whether or not the employee believes an order 
ought to be obeyed. In the sense that the leader must be able to secure 
consent from his followers, it is clear that authority, like power, is re- 
ciprocal. This definition of authority is acceptable if we remember that 
the symbol "consent" has more than merely permissive connotations and 
that consent is only the final, manifest expression of many complex moti- 
vations, mediated by the current interpersonal situation and the personali- 
ties of those concerned. We must ask why the individual accepts authority. 
To say that authority is defined by consent suggests that the subordinate 
has a real choice between acceptance and refusal, that his response to an 
authoritative order is an "either-or" proposition. But this view not only 
neglects the disparities in power and security between the organization 
and the individual and between different individuals in the same organiza- 
tion; it also posits too great a degree of free will and too simple a social 
situation. Here again psychology and anthropology are helpful. As we 
have seen, from infancy on the individual is trained to defer to authority. 
He develops over time a generalized deference to the authority of parent- 
hood, experience, knowledge, power, and status. Moreover, in any given 
dependency situation, many factors operate to negate the "either-or" 
notion. There are so many degrees of compliance, ranging all the way 
from enthusiasm to resignation, that outright rejection of an order be- 
comes a gross and unlikely alternative, particularly among highly so- 
cialized (aggression-repressing) members of the organization. (We as- 
sume that most middle- and upper-level members will tend to be highly 
socialized owing to the technical demands of organization and the ex- 
tended education required to gain the necessary skills.) We assume that 
consent will be normal. In a structured situation, when consent is with- 
held it is expressed in socially acceptable terms: orders are evaded, mis- 
understood, forgotten, or projected upon someone "better qualified," and 
so on. In any event, so long as subordinates know that a superior controls 
ultimate sanctions to compel obedience if his orders are resisted, how can 
authority validly be defined as a matter of consent? 

Theory of Organizational Behavior ! a \ 

When authority and the symbols that define it are organized and patent 
and there are known sanctions to encourage desired reactions, we seem 
to have left the permissive level of influence for the authoritative level 
of power. In terms of a continuum of sanctions, we can say that authority 
is a condition that is subject to being reinforced by sanctions, while in- 
fluence usually secures compliance without reference to sanctions. This 
difference accounts in part for our assumption that interpersonal relations 
in big organizations tend to rest upon authoritative premises rather than 
upon influential ones. This is not to say that organizations do not use 
influence. With the possible exception of those in military organizations, 
interpersonal relations are usually articulated in permissive terms, but there 
is little reason to suppose that those concerned are unaware of the relative 
power disequilibrium. Moreover, aside from other motives for consent 
such as personal ambition, anxiety reduction, and the desire for group ap- 
proval, the very fact that an order emanates from someone with higher 
status and more power tends to induce consent based upon an assumed 
legitimacy of his role. This is only another way of saying that the very 
fact of hierarchy in complex organizations encourages compliance. 

Obviously authority in organizations does not always function hi- 
erarchically. In addition to informal loci of power, there is the fact that 
technical skill demands recognition. Thus a superior must often defer 
where technical considerations are decisive. Yet as the experience of the 
atomic physicists suggests, the control of technical personnel in terms of 
recruitment, promotions, security, and the ends to which the product will 
be put is usually determined by authority according to the formal hi- 
erarchy. The conflicts between administrators and scientists and research- 
ers on this score are well documented. 

An appraisal of authority must also include the fact that big organiza- 
tions are composed of many subhierarchies, each bound together by au- 
thority, interest, and values in a way similar to that in the total organiza- 
tion. Each has its internal power structure headed by a leader who is 
supreme within his own system, but who is a nonleader when viewed from 
the perspective of the larger hierarchy. This devolution of power has im- 
portant consequences. It ensures discipline, since the life chances of those 
in each subhierarchy are determined largely through representations made 
on their behalf by such subleaders. As a result, an upward-looking posture 
characterizes the whole organization. The will of the minority is trans- 
mitted downward through the organization by the subleaders, reinforcing 
their own authority and status vis-a-vis their subhierarchy. 

Here the ambiguity of personal and organizational goals may be seen. 
To retain his position and preserve the hope of future rewards, each sub- 
leader must simultaneously promote organization-wide values and yet re- 
tain the loyalty of his immediate associates by defending their interests 
against both competing subhierarchies and neglect by the elite. Although 

j Theoretical Aspects of Control 

ambivalence may result, his career is in the hands of the elite, and we can 
assume that he will give priority to its will, as he must if he is to fulfill his 
role as an agent for carrying out its policy. He will be measured by the 
loyalty and affirmation with which organization policies are effected. 
Thus the tribute that the upward mobile subleader pays for marginal 
power and localized status is upward-directed anxiety and ambiguous in- 
terpersonal relations. 

In sum, authority includes legal, psychological, moral, and technical 
factors. Their relative weights vary with the particular situation, mainly 
in terms of how manifest and compelling the authoritative stimuli are. 
Thus the hypothesis: The more obvious and powerful {structured) the 
stimuli in a given interpersonal situation are, the more predictable and 
constant the response. 

Despite the complexity resulting from the interplay of these situational 
and psychological factors, we can assume that bureaucratic structure pro- 
duces exceptional probabilities that individuals will defer. We know that 
the recognition that one occupies a hierarchical position clearly sub- 
ordinate to others encourages deference. We also know that a positive 
assumption of authority on whatever grounds enhances compliance. Hay- 
thorn found that "when one member of a group was aggressive, self- 
confident, interested in an individual solution to a task, and showed initia- 
tive, the other members of the group showed less of such behavior than 
they normally did." 11 This tendency reflected the group's desire to avoid 
conflict. The way that group values are imposed will be considered pres- 
ently, after we have seen how the organization encourages the loyalty 
and obedience of its members. 

As Donald Calhoun has suggested, this is done by convincing the fol- 
lowers of the legitimacy and rationality of the organization, mainly by 
equating authority with ethical and ideological principles. 12 Of course, 
all institutions strive to find some basis other than sheer power for their 
authority. Evocative symbols and rituals are enlisted to inspire loyalty to 
the organization. If loyalty is to be thought merited, the values and mo- 
tives, as well as the routine behaviors, of the organization must be seen 
as selfless; if possible the organization must appear to be the embodiment 
of certain universal ideals that are beyond individual criticism. This 
process may be called legitimation. 

Max Weber posited three kinds of legitimacy: legal, traditional, and 
charismatic. The first is based upon the assumption that the organization 
seeks the good of everyone and merits support accordingly. Traditional 

11 Cited in L. F. Carter, "Leadership and Small-Group Behavior," in M. Sherif 
and M. O. Wilson, eds., Group Relations at the Crossroads (Norman, Okla., 1953), 
p. 273. 

12 I am indebted here and elsewhere to his insightful paper, "The Illusion of Ra- 
tionality," in R. Taylor, ed., Life, Language, Law: Essays in Honor of Arthur Bentley 
(Yellow Springs, O., 1957). 

Theory of Organizational Behavior 143 

legitimacy is the belief that the organization and its values are hallowed by 
age and experience and ought not be challenged by any time-blinded in- 
dividual. Charismatic legitimacy is based upon an irrational faith in the 
values and goals of the organization and its leaders. The charismatic per- 
sonality is able to inspire among his followers a desire for sacrifice and 

Most organizations enlist all of these legitimations in justifying their 
claims to loyalty, and the appeals are usually articulated in terms of the 
general welfare. It is necessary, however, for organizations to simplify 
what is really happening, since their objectives are actually more complex 
and less disinterested than this. While they do in part seek to advance the 
common good, they are also concerned with perpetuating the organization 
and its individual prerogatives and with mediating conflicting demands 
within the organization. As Calhoun says, however, if mass loyalty is to be 
maintained, all three activities must be rationalized in terms of the first 
objective. Since it is impossible to define the general welfare, much less 
achieve it, the organization is obliged to draw upon another ideological re- 
source, the myth that it is founded upon unquestionable, unchanging 

Once these principles are accepted, it becomes possible to attribute any 
patent shortcomings, blunders, and injustice of the organization to its 
members, leaving its ideals intact. This sacrificial behavior is seen in the 
dramatic "confessions" that occur periodically in the Communist party, 
but mechanisms that differ mainly in degree are employed by most big 
organizations. Necessity demands that failure be personalized and pro- 
jected in a way which shows that human error was involved rather than 
organizational legitimacy. Certain highly self-conscious organizations, 
among which one can safely include the Marine Corps, the medical pro- 
fession, and the Foreign Service, exhibit this collective idealization, often 
evoking exceptional loyalty from enchanted members. By contrast the 
individual may appear to himself to be ineffectual. The ritual, continuity, 
and power of the organization reinforce this self-perception. 

Another psychological aspect of big organization is the illusion of 
unanimity among its members. Differences of interest and opinion are 
ignored in an effort to present a public image of discipline and unity that 
will enhance the organization's competitive chances. Dissent is confined 
within the organization. Once a decision has been hammered out, everyone 
must accept it, since further discussion L would impair the desired solidarity. 
In part, the tendency of organizations to limit participation reflects a 
desire to avoid the appearance of internal disharmony that active participa- 
tion entails. The common organizational requirement that communications 
be cleared through a "public information" agency is germane. The re- 
marks of uninstructed members can thus be dismissed more easily as un- 
authorized. It follows that only certain individuals are actually responsible 

144 Theoretical Aspects of Control 

spokesmen. These are its priests, who explain the organization to the out- 
side world, interpret its catechism, and rationalize any disparity between 
its ideals of service and its daily behavior. 

To increase the probability that individual behavior will reflect the 
unanimity principle, various appeals are invoked. Affirmative stimuli in- 
clude inspirational calls for loyalty, sacrifice, perpetuation of the organi- 
zation's ideals, and so forth. Negative stimuli are latent but powerful. The 
organization depends mainly upon the sensitivity, the learned deference 
behaviors, the anxiety-reduction needs, and the ambition of the individual. 
Such psychological mechanisms reduce the need for sanctions. This 
climate permits us to view the complex organization as an institution of 
learning which calls upon deep-seated individual needs and experiences to 
support ends that in point of time and significance are often prior to those 
of any given individual. In effect the individual's "self -system" of success- 
ful accommodation to authority is co-opted by the organization, and the 
stimuli that initially induced its development are systematically reinforced 
in ways described below. 

Some of the implications of learning theory for organized behavior can 
now be specified. We suggested earlier that organizations elicited an ex- 
ceptionally strong tendency to defer and, generally speaking, that it 
seemed reasonable to assume that the consistency of individual responses 
was correlated with the power and ease of recognition of stimuli. Psycho- 
logically speaking, the very definition of a "structured field" is that stimuli 
are stable, obvious, and compelling, in the sense that they define appropri- 
ate behavior. Learning is a function of perception and motivation, and 
both depend upon the quality and the number of stimuli, as well as upon 
individual sensitivity and receptivity. We have also defined bureaucratic 
structure as a system of manifest, authoritative stimuli, reinforced by 
known sanctions and a high reward potential. A related hypothesis fol- 
lows: Individual responses will be more certain and constant in bureau- 
cratic structure than in so-called u voluntary associations" To put it an- 
other way, big organizations have decided influence on individual behavior 
patterns, which are defined as a consistent way of reacting to interpersonal 

Perhaps the most common kind of manifest and authoritative stimuli 
are status and prestige indexes. In organizations status and authority are 
designated by appropriate symbols including title, size of office, accessibil- 
ity, and income. In this sense the organization presents a "structured field." 
Such indexes, which differentiate members on the bases of authority, 
prestige, skill, and seniority, enhance the structured character of organiza- 
tions by providing a network of signals that curtail spontaneity, limit al- 
ternatives, and generally define interpersonal relations. For this reason 

Theory of Organizational Behavior 145 

status consciousness tends to become a built-in part of bureaucratic psy- 
chology, as well as a necessary personal skill. 

The present importance of status symbols seems to reflect a change from 
an economy of scarcity to one of conspicuous consumption. But con- 
spicuous consumption is difficult to achieve today because mass produc- 
tion and productivity have made the symbols of material success available 
on so large a scale. The resulting disenchantment of once-favored classes 
is interestingly seen in the Middle East, where the periphery of material 
benefits is slowly being expanded through industrialization and inflation. 
There the elite feels deprived, owing to the loss of indexes that once set it 
apart. In the United States the diminution of this psychic income suggests 
that subtle, nonmaterial distinctions will become more highly valued, since 
they will be more difficult to establish. 

The social framework of status also includes the fact that its symbols 
tend to become a substitute for values no longer attainable. The declining 
opportunity for individual autonomy through self-employment, which re- 
flects the trend toward size, concentration, and difficulty of entry; the 
employment of the "independent" professions on a salary basis; the devalu- 
ation of professional training and increased status anxiety— all seem re- 
lated. The effort to achieve status through word magic is suggested in the 
attempts to borrow prestige by assigning status-laden titles to socially 
devalued jobs: news analyst for reporter, mortician for undertaker, execu- 
tive for salesman, engineer for all sorts of routine jobs, and the widespread 
co-optation of the honored symbol "professional." 

The American assumption of upward mobility, generation by genera- 
tion, is thus related to status idealization. A cross-cultural comparison with 
class-bound European and Middle East societies suggests that in time 
sheer age, the maturing of the economy, and declining occupational mobil- 
ity in the United States will tend to aggravate status consciousness, result- 
ing in greater reliance upon inherited distinctions as objective means to 
status become more difficult to achieve. A free and easy democracy re- 
quires a social and economic situation in which there is relatively free 
access to abundant natural resources. A mature society checkmates this 
competitive situation as the lessons of power are learned by previously 
disadvantaged groups, and an uneasy equilibrium between major interests 
tends to follow. In this milieu big organizations turn to subtle status re- 
wards as compensation for personal dependence and limited mobility. The 
honoring of seniority is an obvious example, as is the small gap between 
initial and upper-level incomes in the bureaucratized occupations. 

As a rule those who have organizational power possess exceptional status 
and prestige reinforcements, such as size and decor of office, expense ac- 
counts, and staff and secretarial assistance (ideally including a handsome 

146 Theoretical Aspects of Control 

private secretary whose loyalty and maternal protectiveness may achieve 
Freudian intensity), that formalize access and encourage attitudes of defer- 
ence. Such stimuli are patent and compelling, and we can assume that the 
responses to them will be more predictable than in less structured situa- 
tions. Their effectiveness is increased by the fact that status anxiety is 
common in big organizations. 

In addition to being obvious and authoritative, organizational stimuli 
are relatively constant. Authority and its symbols are structured so that 
the individuals who personify it may change, but the system of authority 
relationships remains. Indeed, bureaucratic structure may be defined as a 
relatively permanent system of authority relationships. As a result there is 
little ambiguity or uncertainty about rights and obligations which attach 
to the "position" rather than to its incumbent. In comparison with social 
and political power, which is often vague and transitory, organizational 
power is obvious and definable. Moreover, insofar as organized behavior 
is group behavior, the authority of legitimated stimuli is increased by 
sheer numbers. The acceptance of organization values by the majority 
fosters a consensus that makes dissent seem quixotic. 

As was discussed earlier in this paper, such conformity responses have 
a basis in individual learning and experience, namely, in the successive 
authority relationships that begin in childhood. It may be assumed that 
the individual develops considerable sensitivity to authority in all inter- 
personal situations. An example of the resulting pattern of anticipated 
reactions is the effect of rank insignia in the military. The mere sight of a 
high-rank symbol, identifiable at twenty paces, evokes a whole battery of 
conditioned responses from those affected. The relationship is reciprocal; 
all concerned know what their proper roles are. Deference, degree of 
familiarity, tone of voice— indeed the whole character of the interpersonal 
situation is mediated with ease and dispatch by this single evocative cue. 
For most organizations the operational consequence of such signals seems 
clear: the more patent and authoritative the stimulus, the more prompt 
and certain the response. Related functional aspects of status systems in- 
clude the recognition of individual achievement and the legitimation of 
formal authority. 

An interesting latent consequence of status-directed behavior is an ex- 
aggerated picture of conformity demands, which is often dysfunctional 
because it aggravates the fear of action and responsibility often seen in 
big organizations. As A. K. Davis shows, the military situation encourages 
an "affirm and conform" pattern of accommodation, reflecting the over- 
emphasis on authority and status anxiety. 13 This distorted perception re- 

13 A. K. Davis, "Bureaucratic Patterns in the Navy Officer Corps," Social Forces, 
27 (Dec. 1948), 143-53. 

Theory of Organizational Behavior 147 

fleets the anxiety of the individual to please his superiors. Since the elite 
is remote and its will in specific instances cannot be known definitely, the 
individual seeks to anticipate its expectations. As a result such expectations 
may seem more compelling than they really are. The individual is not in- 
clined to underestimate them for fear of alienating those upon whom his 
career chances rest, and he thereby increases the burden of his anxiety. 
This rule of exaggerated response seems to be a common dysfunctional 
consequence of big organization. 

Despite such consequences, the over-all psychological situation is gen- 
erally economical, ensuring internal discipline, dispatch, and a minimum of 
overt interpersonal conflict. Bureaucracy's task is simplified because the 
reactions it evokes are already deep-seated, having been inculcated by a 
succession of social institutions. Since birth the individual has been con- 
ditioned to operate in a structured environment. Noncoercive sanctions 
including custom, mobility expectations, and potential rewards practically 
eliminate the use of gross instruments of control. Because obedience be- 
comes almost automatic, its significance is easily overlooked, or it may be 
repressed as an uncomfortable reality in a society where individualism is 
a pervasive theme. This notwithstanding, authority relations in any society 
become institutionalized between parent and child, teacher and student, 
leader and follower, officer and man, boss and worker, and so on. Al- 
though the resulting power situation may be activated by imperative cues, 
usually the mere presence of an authority figure, his spoken name, or an 
appropriate stimulus such as a title or military rank is sufficient to provoke 
desired responses. We have seen that organisms become conditioned to 
whole classes of stimuli. The patterns of obedience initiated by parents 
become generalized to accommodate a whole range of such authoritative 

We can assume that the anxiety evoked by authority sharpens the in- 
dividual's perception of organization cues. Pavlov was among the first to 
argue (on the basis of empirical observation in mental institutions) that 
anxious people acquire conditioned responses with exceptional rapidity 
and stability. More recent evidence supports this view. Eysenck cites a 
study in which normal individuals required twenty-five repetitions of a 
nonsense syllable accompanied by a buzzer stimulus before a conditioned 
response was established, whereas anxiety neurotics required only eight 
repetitions. 14 Similarly, a study by Franks comparing neurotics, normals, 
and hysterics found that conditioning was much faster and more efficient 
among neurotics. If, as Sullivan insists, most of us spend much of our time 
trying to reduce the anxiety we already have and to avoid getting more, 

14 E. Eysenck, The Psychology of Politics (London, 1954), pp. 260-61. 

! ^8 Theoretical Aspects of Control 

we can assume that anxiety reduction by deferring to authoritative others 
will be a common behavior in complex organizations. Because the range of 
potential responses is thus limited, behavior becomes more predictable. 

Since complex organizations are composed of many small groups repre- 
senting different skills and values, the structure and the psychology of 
such groups must be considered. For this analysis a selected aspect of 
group theory seems most useful. While small groups serve as instruments 
for mediating idiosyncratic personal needs and for wielding informal 
power, they also contribute to the "structured field" being examined here. 
It is well known that after an initial exploratory period small groups be- 
come stratified; authority becomes structured in a way similar to that in 
the larger organization. Freud argued that small-group behavior was best 
understood as an extension of the early family situation in which the 
father's role of authority was assumed by the group leader. We can safely 
conclude that each group develops its own social structure and its means 
of controlling its members. 15 

From the perspective of a given individual, his own group tends to be- 
come "the organization." Usually he performs a given task in company 
with other specialists, organized in a hierarchy on the bases of skill, senior- 
ity, empathy, physical strength, or whatever the going indexes of evalua- 
tion are. Since his work satisfactions and his life chances are often bound 
up with this group, the individual may develop considerable loyalty to it, 
regarding other groups as competitors. He will probably form personal 
alliances within the group, and he will tend to rank each member. Some 
individuals will be accorded leadership, while others will be catalogued as 
followers. The point of reference for such ranking is often the individual's 
perception of his own status in the group, as well as the degree to which 
each member seems to have internalized the group's norms. 

This process of structuring gives the small group a hierarchical quality, 
validated by the tacit endorsement of the entire group. The members ap- 
parently seek an equilibrium so that anticipated reactions become the 
basis of interaction. If groups are to function, such structuring must exist. 
In any informal group situation, once a goal is set certain individuals 

15 Among others, see W. F. Whyte, Street Corner Society, 2d ed. (Chicago, 1955), 
and his "Status in the Kitchen," in Human Relations in the Restaurant Industry 
(New York, 1948); George C. Homans, The Human Group (New York, 1950); 
L. Festinger, S. Schacter, and Kurt Back, Social Pressures in Informal Groups (New 
York, 1950), chs. v, vi; A. Hare, F. Borgatta, R. Bales, Small Groups: Studies in 
Social Interaction (New York, 1955); Peter M. Blau, The Dynamics of Bureaucracy: 
A Study of Interpersonal Relations in Two Government Agencies (Chicago, 1955), 
ch. x. 

Theory of Organizational Behavior 149 

gradually assume positions of leadership; in a relatively brief time the re- 
sulting pattern becomes crystallized because it meets both operational and 
emotional needs. In such situations the individual tends to seek "consensual 
validation" of his own attitudes by comparing them with those of the 
group. That is, he tends to look elsewhere, to the group and its authority 
figures, for cues that define approved opinions. 

The individual's deference to group or majority norms has been estab- 
lished by many experimental studies. 

A study by L. Coch and J. R. P. French dealing with resistance to 
change in a factory is also suggestive. 16 Following a slight change in her 
job, a machine operator produced at a reduced rate of fifty units per hour. 
Some ten days later, however, she again reached normal production and 
soon began to exceed the rates of her group. She then became the target 
of considerable abuse as a "rate buster," whereupon her productivity de- 
creased to the group's level. Three weeks after the change, all the other 
operators were transferred, leaving this girl alone. Within four days she 
was turning out eighty-three units per hour, and she produced steadily at 
that rate thereafter. 

Compliance in organizations is thus encouraged by a variety of sanc- 
tions, most of which invoke the anxiety-conformity-approval syndrome 
but vary considerably according to the situation and the personalities con- 
cerned. Given dominant values of success and security, middle-class child 
training and education seem to foster a high degree of adaptive anxiety, 
discipline, and repression of aggression in outside-the-home interpersonal 
relations, whereas a lower-class milieu is somewhat more tolerant of ag- 
gression. In industrial work situations ridicule, censure, and even blows 
are used to discipline nonconformants. On the other hand, in organiza- 
tions engaged in highly technical work requiring considerable education 
and training (correlated in turn with middle-class social expectations), we 
find that sanctions are apt to be rather more Machiavellian and that re- 
wards meet status needs to a greater extent than they do economic ones. 

Any useful theory must account for such differences, and research in 
the framework outlined above would have to differentiate among organi- 
zation members according to class, motivation, educational background, 
and attitudes toward authority, since these factors play a significant role 
in accommodation. I would propose three general patterns of accommoda- 
tion to the bureaucratic situation: the upward-mobiles, the indifferents, 
and the ambivalents. (There is some evidence that items in the Adorno 

16 L. Coch and J. R. P. French, "Overcoming Resistance to Change," Human Rela- 
tions, 1 (1948), 512-32. 

1 50 Theoretical Aspects of Control 

"F" scale are helpful in identifying each type. 17 ) A preliminary sketch of 
each ideal type follows. Type one is characterized by an ability to identify 
with the long-range, abstract goals of the total organization and to make 
these a meaningful basis for participation, in other words, to accept the 
legitimacy and rationality of the organization. Allied with this attitude is 
a capacity for action despite conflicting alternatives and contradictory 
aims; the organization's values are accepted as decisive. An acceptance of 
the demands and operational necessity of the organization's authority and 
status systems seems another functional attitude. For example, it is well 
known that the successful executive tends to regard his superiors as 
friendly and sympathetic. 18 These formulations permit us to suggest a 
third major research hypothesis: Individual patterns of accommodation 
to the organization are associated with attitudes toward authority and 
with socioeconomic status. 

Type two, the indirTerents, seem to comprise the most common pattern 
of accommodation. Rejecting majority values of success and power, the 
indifferent's orientation is essentially extravocational. His work is sepa- 
rated from the assumed-to-be more meaningful aspects of life. His refer- 
ences lie outside the organization, which merely provides the income 
necessary to indulge ofT-the-job satisfactions; and unlike the upward- 
mobiles, these activities rarely reinforce his organizational role. His rela- 
tionship with the organization is essentially an economic bargain in which 
he sells his time and energy for a certain number of hours per week but 
jealously guards the remaining time as his own. 

The third type of adjustment pattern is to be seen in the ambivalents, 
who comprise that small minority who can neither resist the appeals of 
power and success available through the organization nor play the role re- 
quired to attain them. The ambivalent seems to need security, which the 
organization's structure and power could provide, but he is tempera- 
mentally unable to make the accommodation necessary to obtain it. This 
conflict seems to reflect inapposite views toward authority and an aggres- 
sive sense of individuality which will not permit him to accept the or- 
ganization as a collective instrument seeking ends that are beyond those 
of any individual in point of time and significance. He is thus unable to 
make decisions in terms of organizational premises, exhibiting instead a 
particularistic point of view which places friendship and similar subjective 
values above the objective universalistic values that ensure success in the 
upward-mobiles' case. 

17 T. W. Adorno et al. y The Authoritarian Personality (New York, 1950). 

18 Among others, see Burleigh Gardner, "Successful and Unsuccessful Executives,' 
Advanced Management, 1 (Sept. 1948), 116-25. 

General Theory of Administration i 5 1 


Edward H. Litchfield* 

Dean Edward H. Litchfield offers several major and minor propositions 
•which may provide the beginnings of a framework for a general theory 
of administrative action. 

I hope that this essay will add several propositions which may bring us 
a little closer to a working theory of the nature of the administrative 
process. I do not believe we have such a theory today. There are urgent 
reasons for the early development of at least a working theory of ad- 
ministration. Three are particularly compelling. 

First, it is virtually impossible to codify our existing knowledge with- 
out some conceptual framework within which to do so. Theory is im- 
portant for this purpose in any field of investigation, but it is crucially 
significant in an applied field which must ultimately draw together 
knowledges now scattered through all of the social and behavioral sciences 
and through the many applied areas of business, public, military, hospital, 
and educational administration. As a framework for the organization of 
materials, a working theory is equally needed by the management con- 
sultant, the teacher, the professionally conscious administrator, and the 
research worker. 

Second, a comprehensive theory is needed as a guide to research. How- 
ever tentative that theory may be, it should help to discern gaps in both 
existing knowledge and ongoing research and thus to further the design 
of other research efforts. It would also provide working hypotheses as 
guides to individual research efforts and as vehicles for the subsequent 
incorporation of research efforts into organized bodies of thought. 

Finally, a tenable theory of administration could become an extremely 
useful guide to administrative behavior. The analytical and intellectually 
self-conscious practitioner will readily recognize the importance of a 
broad theoretical framework which he may use as a measure of his per- 
sonal performance and which may provide a behavioral check list in his 
day-to-day undertakings. Educators who do not subscribe to the "either 
you are born with it or you aren't" school will also find it of primary im- 
portance in shaping curricula and in guiding potential administrators. 

*From Administrative Science Quarterly, I (1956), 3-29. Reprinted by permission 
of the Administrative Science Quarterly. 

! r 2 Theoretical Aspects of Control 

However urgent a general theory of administration may be, it is un- 
likely that it can be set forth in the near future. Certainly this essay does 
not pretend to present a general theory of administration. It will attempt 
to set forth a series of working hypotheses or propositions which may 
provide at least the beginnings of a framework for a general theory of 
administrative action. They will serve their purpose best if they are used 
as targets for future effort. 

FIRST MAJOR PROPOSITION: The administrative process is a cycle 
of action which includes the following specific activities: 

A. Decisionmaking D. Controlling 

B. Programming E. Reappraising 

C. Communicating 

This pattern of actions is found in various forms in all phases of ad- 
ministration. It occurs in policy areas; it is essential to personnel, finance, 
and other types of resources management; and it is to be found in the 
executive function as well. The specific activities and the cycle as a whole 
provide the mechanism by means of which all of the separate functions of 
administration are carried on. It is at once a large cycle which constitutes 
the administrative process as a totality and a series of small cycles which 
provide the means for the performance of specific functions and sub- 
functions and even for individual technical activities. 

In an idealized form it occurs as a logical sequence in which there is a 
progression from the making of a decision to the interpretation of the 
decision in the form of specific programs, to the communication of that 
programmed decision, to the establishment of controls for the realization 
of the decision, and finally to a reappraisal of the decision as programmed, 
communicated, and controlled. In fact, however, the cycle often occurs 
in abbreviated form. Thus the practicalities of programming a decision 
may lead to immediate reappraisal, eliminating the steps of communication 
and control. Again, total group participation in decision making may 
eliminate much of communication. If individual steps are abbreviated or 
even eliminated, the cycle is nonetheless complete. In fact, the steps 
probably are there, even though in quite attenuated form. 

Many such cycles are in action in the administrative process at any one 
time. One elaborate cycle may be proceeding at board of directors' level 
regarding fundamental objectives, while smaller and still sequential cycles 
may be going forward in finance and sales, and at the same time very im- 
mediate, specific, and perhaps abbreviated cyclical actions occur in the 
office or in the mind of a district sales manager concerned with a particu- 
lar problem devoid of any policy, methodological, financial, or human re- 
lations significance. There is thus a series of wheels within wheels, tangent 

General Theory of Administration 1 5 3 

now at one point, now at another. The totality is administrative action, 
and the wheels are similar not in size but in the articulate and inarticulate 
uniformity of their components. 

The grouping of these activities is made cyclical by the presence of the 
activity of reappraisal, for this brings the sequence back to substantially 
the point at which it began. Yet while it completes a full cycle the se- 
quence does not necessarily lead again to identical action. If the original 
decision is precisely reaffirmed, the sequence of the five activities is no 
more than a revolution around a constant axis. However, if the original 
decision is modified in the light of evidence presented in the reappraisal, 
the axis may move and the circle take on a cycloidal form. With the pas- 
sage of time and subsequent revolutions in the cycle, an extensive cy- 
cloidal pattern may develop. 

minor proposition: Decision making may be rational, deliberative, discre- 
tionary, purposive, or it may be irrational, habitual, obligatory, random, 
or any combination thereof. In its rational, deliberative, discretionary , 
and purposive form, it is performed by means of the following sub- 

a. Definition of the issue 

b. Analysis of the existing situation 

c. Calculation and delineation of alternatives 

d. Deliberation 

e. Choice 

The sequence of activity from definition of issue to choice is again 
idealized. It presumes rationality and the existence of discretion. It con- 
templates the opportunity for deliberation, the possibility of calculating 
alternatives, and the existence of knowledge with which to estimate the 
situation. It is seldom that all of these factors are in fact present. Yet only 
if we view the pattern in its idealized form are we able to understand the 
nature of the parts as they occur individually or in combination. 

Definition of the issue is the isolation both of problems (in the sense that 
a problem is a difficulty) and also of opportunities in which no difficulty 
is present. Thus it has both a corrective and creative aspect. A problem 
may require diagnosis, whereas an opportunity may be defined by re- 
search. In any event the function of issue definition is the clarification and 
description of the question at hand. Efficiency in subsequent steps in the 
cycle is obviously dependent upon the precision with which this first 
activity is undertaken. 

Situation analysis involves a systematic effort to present facts regarding 
the existing situation where they may be known and estimates regarding 
that situation when facts are impossible to obtain. This must include a 

j - 4 Theoretical Aspects of Control 

factual statement of prevalent values when those are part of the situation 
and relevant to subsequent choice. Many techniques assist the administra- 
tor in the performance of this activity. They include accounting, opinion 
surveys, market analyses, field testing of products, operations research, in- 
telligence reports, and countless other similar analytical tools. 

Alternative calculation involves two major steps: first, a systematic iso- 
lation and description of known alternative courses of action; and, second, 
a statement of the consequences of the alternatives where the latter are 
known. Where they are unknown, they must be estimated. These estimates 
will often themselves take the form of alternatives. In any event, this dis- 
tinctive activity is concerned with known facts, assumed facts, and factual 
statement of values. Here we are assisted by such methods and techniques 
as economic forecasting, market projections, linear programming, and 
game theory. 

Deliberation is the next step. In it one is concerned with reviewing the 
issue in the light of what is known in the existing situation and with regard 
to the alternative courses of action which appear to be available. It in- 
volves an assessment of values, an appraisal of probabilities where chance 
alone is involved, and strategy where knowledge is imperfect. Delibera- 
tion approaches rationality as the values become explicit, as probabilities 
are analyzed, and as risk calculation can be reduced to a mathematical 

Having defined the issues, assembled and stated the facts regarding the 
existing situation, calculated and delineated the alternative courses of ac- 
tion with their known and estimated consequences, and reviewed all in 
terms of an explicit statement of ordered values, one is then prepared to 
choose. Choice under these circumstances is influenced by several consid- 
erations: first, free will or discretion and, second, the presence of ration- 
ality. Thus, a "wise choice" is apt to be one which the administrator had 
the discretion to make, had the rationality to base upon the known and 
estimated data at hand, and had the critical faculty to appraise in terms of 
the relative significance of those data. 

In making choices, the administrator must understand that there is not 
always one right answer. Truth is frequently plural, and therefore the ob- 
jective in the exercise of choice is rationality and not a pursuit of a non- 
existent absolute. Failure to recognize both the plurality and relativity of 
correctness in decision making may lead to a time-losing indecision and a 
precarious mental health for those who must choose. 

Actually few decisions are made by means of this full sequence of ac- 
tions. The issue may be so patent as to make definition unnecessary. Often 
little effort is made to ascertain or estimate facts. Built-in biases frequently 
result in only the most superficial calculation of alternatives, and delibera- 
tion may be short-circuited by unspecified values or an unwillingness or 

General Theory of Administration 1 55 

inability to think in strategic terms. The elimination of certain of the steps 
may mean poor decisions, or again it may mean that specialized circum- 
stances make one or more steps self-evident or unnecessary. 

We must also recognize the extent to which decision making is in- 
fluenced by limitations upon rationality. The administrator must not only 
allow for his own irrationality, he must also calculate the actions of others 
as being both rational and irrational. Thus, our decisions may be partly 
rational and partly irrational. 

minor proposition: Decisions become guides to action after they have 

been interpreted in the form of specific programs. 

Decisions must be interpreted by specific programs which provide the 
direction for detailed operation. These might be called plans, were "plan- 
ning" not a confused term which is sometimes used with reference to an 
"outline of alternatives," and in this sense is a part of decision making. 
Again, the term is used as synonymous with "programming" as the latter 
term is used here. One may therefore more accurately speak of planning 
alternatives (in the decision-making process) and of program planning as 
an activity designed to implement decisions. Program planning rests on a 
wide range of specific methods and techniques. These include capital 
budgets, operating budgets, manning tables, organization charts, tables of 
equipment, and a variety of similar means of translating a decision into 
specific programs for the allocation of money, manpower, authority, 
physical resources, and so on. The completeness of the program is a de- 
termining consideration in the effectiveness of the original decision. 

minor proposition: The effectiveness of a programmed decision will vary 

with the extent to which it is communicated to those of whom action 

is required. 

Communication follows the programming of decisions in cases in which 
those who must act have not participated in the original decision. Com- 
munication is a method by which an individual or group transmits stimuli 
which modify the behavior of another individual or group. While it is an 
activity employed in administration, it is obviously broader than the ad- 
ministrative process, for there is communication among individuals who 
have no administrative relationship to one another. We may therefore say 
that we are concerned here with the use of the method of communication 
for the restricted purposes of the administrative process. 

Communication in administration involves three primary responsibilities. 
First, the administrator must establish the channels, the methods, and the 
opportunities for communicating with all of those above, below, and 
around him whose actions he would influence. Second, he must establish 

1 56 Theoretical Aspects of Control 

channels and provide the opportunity for others to communicate with 
him. Third, he must assure the existence of channels of communication 
among all those in the organization who must influence one another if the 
organization is to achieve its total objectives. Each of these is a deliberate 
action which he must take. In two instances it is his responsibility to pro- 
vide structure on the one hand and to utilize it on the other. 

minor proposition: Action required by a programmed and communicated 
decision is more nearly assured if standards of performance are estab- 
lished and enforced. 

Standard setting and enforcement may be more generally known as 
"control." Communication was concerned with stimuli which would call 
up desired responses, and control is concerned with a definition of the 
desired response and the methods of assuring its occurrence. In other 
words, control is an action which provides norms which will serve as a 
guide to the actors and against which to measure their actions. Both stand- 
ard setting and enforcement are carried on by means of elaborate tech- 
niques of control. They may be techniques designed to control basic 
programs and operations, such as a budget, an organization chart, or a 
functional statement. Or they may be processive tools such as job stand- 
ardization, wage and salary schedules, purchasing specifications, or cost 
and quality controls. All play essentially the same role in the action cycle. 

A notable characteristic of control action is its tendency to become an 
end in itself. Thus we have the familiar phenomenon of the controller who 
seems more concerned with his accounting mechanisms than with the 
management purpose which they presumably exist to further. This often 
results from the fact that there is an internal element of completeness 
within the control activity. Having set a standard, reviewed performance 
in terms of the standard, and then enforced the standard, the person per- 
forming the action has, in a sense, completed a full and satisfying cycle. 
Standard enforcement in fact becomes standard realization and, hence, 
achievement. It is the only action in the cycle outside of decision making 
which has this organic unity about it which encourages its use as end 
rather than means. 

While the subactions of control are standard setting and enforcement, 
the primary working methods are determined by the properties of the 
thing which is controlled. Thus control of people is achieved by means of 
a skillful manipulation of various types of rewards and punishments de- 
signed to appeal to the several motives of the groups and individuals con- 
cerned. Control of money comes with skillful manipulation of that re- 
source in terms of its own laws of behavior. 

General Theory of Administration 1 57 

minor proposition: Decisions are based on facts, assumptions, and values 
which are subject to change. To retain their validity, decisions must 
therefore be reviewed and revised as rapidly as change occurs. 

A decision which has been programmed, communicated, and controlled 
has validity only for the limited period in which the facts, assumptions, 
and values upon which it was based have retained their original character. 
Only for such a period can the first four steps in the action cycle be re- 
garded as static. In fact, not only are the facts, assumptions, and values in 
a state of constant flux, but the fact of decision in itself often brings sub- 
stantial alteration in the total pattern of circumstance on the basis of which 
the decision was made. Hence a fully articulated decision— that is, one 
which has been made, programmed, communicated, and controlled— in it- 
self brings about sufficient change to necessitate its own reconsideration. 
This is the activity of reappraisal. 

Reappraisal is necessitated not only by change but by the possible im- 
perfection of the original decision which time and circumstance may 
make apparent. New insights may be gained which improve the ad- 
ministrator's understanding of a more nearly correct decision, even though 
the facts themselves have not been altered. Here we return to a recogni- 
tion of the plurality of truth as noted in our discussion of the decision- 
making activity. Thus we reappraise decisions because of our acceptance 
of the concepts of "contingent universe" and "organic incompleteness" 
which are implicit in cybernetic thought. 1 

Reappraisal may be accomplished in several ways. In its simplest form 
it is no more than a review of the original issue in terms of new data, new 
assumptions, new strategies, and new values which have bearing on the 
decision but which arise from extraneous sources. Thus, new information 
about Soviet military production may require the reappraisal of a foreign 
policy decision which had been based upon different fact assumptions. 
This may be referred to as a "feed-in" activity. Quite different is the 
process of "feed-back," which is the essence of cybernetic theory. Here 
we contemplate the reappraisal of an original decision upon the basis of 
facts and values which have been generated by and as a result of the 
original decision. Reappraisal therefore provides for self-generated change 
and growth. 

Whether the reappraisal be "feed-in," "feed-back," or a combination, 
its function is the same. It is needed to complete the action cycle in order 
to make it dynamic rather than static. This is the action which induces 
change and growth. It must be specifically provided for in the action 
pattern if growth is to be accepted as both constant and necessary. Only 

1 See Norbert Wiener, The Human Use of Human Beings (Boston, Mass., 1950). 

1 5 8 Theoretical Aspects of Control 

through reappraisal can administration adjust to the constancy of evolu- 
tion; otherwise the administrator is apt to pursue a concept of the perma- 
nent and absolute. This is a vehicle for incorporating into administration 
an understanding of stability through change rather than through an 
artificial staticity. 

SECOND MAJOR PROPOSITION: The administrative process func- 
tions in the areas of: 

A. Policy B. Resources C. Execution 

A "policy" is a definition of those objectives which guide the actions of 
a whole enterprise or a significant portion thereof. It is thus distinguished 
from the general term "decision," which may guide actions without refer- 
ence to such objectives. The "resources" of administration are four: peo- 
ple, money, authority, and materials. "Execution" is a function of integra- 
tion and synthesis which is intended to achieve a dynamic and total 
organism. All functional areas are requisite to the process. Execution di- 
vorced from policy is aimless. Similarly, the policy function tends to be- 
come remote and sterile unless associated with resources and execution. 

minor proposition: Action in each functional area is accomplished by 
means of the action cycle previously described. 

The policy function is conventionally referred to as "policy making" 
or "policy formulation." In fact it is far more. Policies are not only made, 
they are also programmed, communicated, controlled, and reappraised. It 
is only in this total sweep of the action cycle that the policy function has 
full meaning and is satisfactorily distinguished from the activity of de- 
cision making. 

The action cycle is also the vehicle for the accomplishment of the re- 
sources function. In determining the need for money or people, the ad- 
ministrator defines his problem, estimates his situation, calculates his alter- 
natives, makes a choice, and thus in fact makes a decision. In allocating 
personnel by manning tables, authority by functional statements or charts, 
or money by budgets, he is in fact programming his decision. Direction of 
personnel is communication. Organization charts and manuals, budgets 
and inventories, are forms of control of money, authority, and materials. 
The final step of reappraisal is provided for in budget analysis and revi- 
sion and in a whole series of other resource function techniques. 

The action cycle is repeated in the executive function. Setting the poli- 
cies and resources in motion, synthesizing their conflicting values and 
tendencies, and integrating the resulting management are achieved by a 

General Theory of Administration i 59 

constant series of cyclical movements from decision to reappraisal to new 
decision and further reappraisal. Maintaining them in dynamic equilibrium 
is realized in the same way. 

minor proposition: Each junction seeks a value which when realized is 
its contribution to the administrative process. 

Policy seeks purposive direction for the enterprise. The resource func- 
tion seeks economy both in the sense of productivity and of frugality. 
Execution seeks and is evaluated by the degree to which it achieves a state 
of dynamic coordination. These are the contributions of the three func- 
tional areas. Together they constitute an organism whose direction is 
purposive and whose resources are productively and frugally employed. 

minor proposition: Each junction has distinctive characteristics which 
govern the application oj the cycle to it. 

We have observed that policies involve questions of value and fact. 
Values are plural and facts are contingent. As a consequence the action 
cycle employed in the policy function is modified. The' four resources 
have different characteristics. People are moved by varying combinations 
of rewards and punishments. Money is moved by factors of scarcity. Au- 
thority has properties which influence the way in which it may be al- 
located and exercised. The executive function achieves synthesis and main- 
tains a dynamic organism by observing the laws of equilibrium and decay 
whether they be drawn from modern group dynamics or are as remote as 
Henri Bergson's "law of twofold frenzy." 2 In each case, the cycle is con- 
stant, but it is performed in the context of a function which responds to its 
specialized properties. 

minor proposition: The junctional areas oj administration are integrally 
related to one another. 

We have already observed that each of the functions is requisite to the 
total process. It is equally true that the areas are integrally related to one 
another. Obviously policy is the major determinant of the character of 
the resource and executive functions, but it is also true that resources are 
important determinants of policy and that execution may be either the 
realization or destruction of policy. Administrative behavior must be cal- 
culated to recognize that a new policy has immediate implications for 
authority, finance, and personnel. One follows the other automatically. 
No one can be isolated from the other, for they are in fact a continuum 
of reciprocating parts. 

2 Henri Bergson, The Tivo Sources of Morality and Religion (Garden City, N.Y., 
J 954>- 

! 60 Theoretical Aspects of Control 

THIRD MAJOR PROPOSITION: The administrative process is carried 
on in the context of a larger action system, the dimensions of which 
are: 3 

A. The administrative process 

B. The individual performing the administrative process 

C. The total enterprise within which the individual performs the 

D. The ecology within which the individual and the enterprise 

Each of these dimensions has both constant properties and variables. 
We have examined the constants of the administrative process and must 
now note its variables. Likewise, we must observe the way in which 
these dimensions are related. In fact, the process which we have examined 
in the abstract becomes a real thing only in the hands of the persons per- 
forming it and in the context of specific total organizations and, in par- 
ticular, total environments. These three dimensions affect the administra- 
tive process by altering its variables. It in turn alters each of them in its 
impact upon their variables. We thus have a concept of a system contain- 
ing four dimensions, each of which has a structure comprising a number 
of variables which interact upon one another. 

Furthermore, the other dimensions, have no constant impact on the ad- 
ministrative process. At one point the impact of the individual administra- 
tor may be decisive and at another time relatively inconsequential. Thus 
in a highly articulated bureaucracy, the variations among administrators 
will affect the way in which the process is performed to lesser extent 
than in a new organization which has been less rigidly structured. In 
other words, there are not only variables within each dimension but there 
is variation in the relative roles among the dimensions in this total action 

minor proposition: While constant in basic structure, the administrative 
process will vary in important aspects, depending upon the personality 
of the person performing it. 

The cycle of administrative action and the functions of the administra- 
tive process are constants regardless of who performs them. The manner 
in which the actions are taken and the functions are accomplished, how- 
ever, will vary with the characteristics of the individual. These variations 
in manner are as important as the constancy in structure. The self-con- 
tained administrator may deliberate alone, while the new and uncertain 

executive may take elaborate counsel. Deliberation is present in both 

3 Here the term "dimension" refers to a category of variables. 

General Theory of Administration 1 6 1 

cases, but the methods of its exercise are importantly different. Or, in 
estimating the existing situation, a Wilson may collect his own facts, 
whereas an Eisenhower will assemble information by means of an elabo- 
rate staff organization. Different individual administrators have radically 
different effects upon the whole organization. Authority as a resource may 
be delegated to many by a generalist, or to a few by a specialist. As a 
consequence the authority structure will be flat in one case and pyramided 
in the other. The consequences of these varying ways in which two types 
of personalities allocate authority are far reaching, but the resource func- 
tion has nevertheless been satisfied. 

minor proposition: While constant in basic structure, the administrative 
process will vary in important respects, depending upon the character 
of the total enterprise within which it is performed. 

The way in which the administrative process is performed will vary 
with the character of the organization. One-man decision making in an 
academic atmosphere is less likely than in a family-owned manufacturing 
organization, yet there is decision making in both cases. Reappraisal may 
be infrequent in a conservative British textile firm but constant in a 
young and aggressive corporation like General Dynamics. 

The administrator communicates in a different way in an organization 
with a well-developed, informal structure than he does in an enterprise 
which has a high turnover rate and which is composed of persons whose 
backgrounds and associations are quite diverse. Yet the communication 
activity is there; only the ways of its exercise and the degree of its effec- 
tiveness will change. The impersonality of social relations which increases 
as an organization grows in size and complexity presents problems in 
communication unknown in simpler organizations. 

minor proposition: While constant in basic structure, the administrative 
process will vary in important respects depending upon the environ- 
ment in which the individual and total enterprise function. 

The administrative process will also vary with the physical, cultural, 
and technological environment within which it is performed. Communica- 
tion is obviously influenced by a changing technology which eliminates 
much of the significance of distance. It is similarly influenced by conversa- 
tional practice resulting from social systems of rank and class. Effective- 
ness of the control activity may depend upon the financial and psycholog- 
ical resources which the community provides as alternatives to acceptance 
of a distasteful standard of working norms. Indonesian understanding of 
"the good neighbor" raises problems in limiting supervisory spans which 
are missing in societies where efficiency means more and neighborliness 
less. Controls which may be imposed and enforced in one atmosphere may 

j 5 2 . Theoretical Aspects of Control 

be vitiated by the existence of community or professional standards or 
values which preclude the individual's accepting the control provided by 
administration. Part of the theory of the Nurnberg trials is closely related 
to this. 

minor proposition: The types of relationships existing among the three 
dimensions other than the administrative have an effect upon the ad- 
ministrative process. 

My colleague, James Thompson, has suggested that "there appear to be 
four major types of relationships which an enterprise may have with or- 
ganized elements of environment." 4 He notes competition, bargaining, 
cooptation, and coalition as the primary types of relationships between 
the dimension of the whole enterprise and the dimension of the environ- 
ment. He then points out that these relationships between two dimensions 
have important bearing on the decision-making activity, for they alter 
the way in which it is carried on and vary the number of the participants 
therein. Thus the relationships between two dimensions have caused varia- 
tion in the third, the administrative process. It would appear probable that 
we may generalize beyond this and say that there are definable relation- 
ships among each of the three nonadministrative dimensions and that the 
variations among those relationships will have consequent bearing upon 
the administrative process itself. 

FOURTH A4AJOR PROPOSITION: Administration is the performance 

of the administrative process by an individual or a group in the context 

of an enterprise functioning in its environment. 

The administrative process is a series of interdependent steps which may 
be isolated and described in the abstract. Administration, on the other 
hand, is the performance of the process in the specific contexts of enter- 
prise and environment. As such it is primarily behavior, though in other 
times and cultures it may have been thought of as largely law. 

We have already observed that the parts of the action cycle are re- 
ciprocally influential, that the functions performed by the administrative 
process are integrally related to one another. They suggest an interde- 
pendence which is increased once the interdimensional influences have 
been introduced. 

This complex of interrelationships probably constitutes a whole, though 
it is not yet clear whether this entity can be referred to as a "system" with 
all of the "organic" implications of that term as it is used in the life sci- 
ences. There would, however, appear to be a totality in administration 
(not in the administrative process) which is significant. 

4 "Administrative Process Working Papers" (unpublished Cornell manuscript, Dec. 
3i. 1955)- 

General Theory of Administration 1 6 3 

minor proposition: Administration as a totality has definable attributes. 
They are: 

a. It seeks to perpetuate itself. As a definable complex, administration 
has many of the characteristics of other total organizations. Like them, its 
first attribute is its tendency to perpetuate itself. 

b. It seeks to preserve its internal well-being. It is sufficiently self- 
conscious to attempt to protect itself against disruption from within or 
destruction from without. It is therefore concerned with its own internal 
workings and the morale and welfare of its participants. 

c. It seeks to preserve itself vis-a-vis others. Each individual behavior 
pattern composing a complete administrative process maintains a competi- 
tive relationship with other behavior patterns constituting other processes. 

d. It seeks growth. Like all other organisms, it is aware of the fact that 
it cannot stand still for long but must go either forward or backward. In 
its dynamic phases, therefore, it normally seeks to grow. Much of the im- 
petus for mergers, for "empire building," and for entrepreneurial effort 
in general is made not only on behalf of the corporation, or the bureau, or 
the institution, but also on behalf of a separate and identifiable administra- 


minor proposition: The attributes of the totality of administration have 

significant effects upon administrative behavior. 

These attributes of totality are really properties of organic compulsion 
and as such are compulsive as far as behavior is concerned. Thus, the suc- 
cessful administrator seeks internal cohesion among the members of the 
administrative group. He seeks to keep his group intact. He presses for 
individual identification with the total process. He stresses competition 
as a means of preserving his "administration's" relationship to others. In 
short, he performs in such a way as to attempt to perpetuate the process, 
maintain its internal well-being, preserve its position among competitors, 
and, indeed, to help it grow. 

FIFTH MAJOR PROPOSITION: Administration and the administrative 
process occur in substantially the same generalized form in industrial, 
commercial, civil, educational, military, and hospital organizations. 
The concept of the universality of administration and of the administra- 
tive process has been implicit in much which has been set forth above. It 
must now be made explicit as a separate proposition. This is particularly 
important both for the classification of existing knowledge and as a hy- 
pothesis for subsequent investigation. 

The cyclical development of administrative action, beginning with 
decision making and moving through reappraisal, occurs in all types of 
organizations. Similarly, each of them is served by administration through 

!6 4 

Theoretical Aspects of Control 

the accomplishment of the same basic functions in the areas of policy, re- 
sources, and execution. Again, the process is no less organic in a hospital 
than it is in a manufacturing establishment. Finally, the process is but a 
portion of a larger action system, whether that system occurs in the De- 
partment of the Interior or in General Motors. The process becomes a 
whole as administration when performed by an individual within an enter- 
prise functioning in its own ecological setting. 

In every case, there is a constancy in fundamental properties. The dif- 
ferences which exist from one field of application to another are differ- 
ences which result from the factors suggested in our discussion of the four 
dimensions constituting action. These are the fundamental differences; 
the variations in institutional application are derivative. When thus an- 
alyzed, however, these more fundamental differences are seen to be but 
variations in the way in which a constant process is performed or accom- 
plished. They do not argue against a basic universality. 

Section B: 


The administrator of a business enterprise has aggregate 
administrative responsibilities and specific areas of control. 
These areas are cash, inventory, research and development, 
production processes, marketing operations, and personnel. 
It is necessary for the administrator to exercise managerial 
control over these specific areas to manage the enterprise 

In so doing the executive must recognize that the impor- 
tance of cash forecasting in forward planning ties in the 
establishment of a continuous balance between the cash 
position and the other aspects of the plan. The short-term 
cash plan is measurable against recent actual performance 
and current trends, while the long-term cash plan is subject 
to such broad influences as national economic trends, the 
effect of technological and competitive progress on cus- 
tomer demand, and the development of new products and 
new markets. These forces direct management into continual 
reappraisal of manufacturing facilities, methods of procur- 
ing materials, labor supply, and future cost-selling price 

When cash budgets are prepared some firms also draw up 
an application of funds statement. The preparation of both 
statements helps to spotlight trends through the cash budget 


balance sheet. The fundamental purpose of the funds state- 
ment is not to tell more about the income-generating and 
income-measuring processes, but rather to disclose data 
about the related but distinct task of financial management 
of the business enterprise. 

The inventory management objective of the enterprise 
should be to maintain balanced, strategically located stocks 
of inventory available for quick shipment. A concerted pro- 
gram will be required on the part of general, sales, financial, 
purchasing, and manufacturing management since all will 
have the same vital objective of customer good will and the 
addition of new customers who are attracted by progressive 
inventory management and reliability of source, along with 
factors of price and quality. The decision and control of 
finished goods should generally be assigned to the merchan- 
dising group, since they are most familiar with customers 
and their requirements. 

One of the major tests of a good inventory control and 
production planning system is its ability to meet customer 
requirements based on good sales forecasting and good pro- 
curement and manufacturing practices. Usually a company 
having a high degree of reliability in meeting shipping prom- 
ises will also have a high profit within the industry. 

It is possible to have inventory controls which are not 
only flexible but also carefully designed and explicit. The 
task needs special analytical tools; in a complicated business 
it defies common sense and simple arithmetic. Methods must 
be employed to take direct account of uncertainty and to 
measure the response characteristics of the system and relate 
them to costs. Such methods are the distinctive mark of a 
really modern, progressive control system. It takes a good 
deal of research into sales and product characteristics, plus 
skill in sensing which of the many possible approaches are 
likely to be fruitful. 

Some businessmen are prone to view inventories with dis- 
taste, as an apparently necessary drain on resources, some- 
thing that no one has been able to eliminate but hardly a 
"productive" asset like a new machine or tool. Actually, 
however, inventories are as productive of earnings as other 
types of capital investment; they serve as the lubrication 
and springing for a production-distribution system. A com- 
prehensive inventory control system should be closely co- 


ordinated with the other planning and control activities, 
such as sales forecasting, cash planning, and capital budget- 
ing, since it affects all of these activities in many ways. The 
specific steps and timing will vary from one company to 
another, depending on product and process requirements, 
but the essentials of an inventory control system can be 
grouped into three broad classes: long-range inventory plan- 
ning, short-range planning, and scheduling. 

Planning is the key to administrative control. At the com- 
pany level, it determines the amount of money that can be 
allocated to research. The research management distributes 
these budget dollars to fit the particular program. The cost- 
reporting procedure then gives the management tools to 
analyze, control, and schedule research activities to the best 
future interests of the company. 

The establishment of a budget does not give the directors 
of the research and development program any guide for 
programming their efforts. Such guidance is invaluable for 
those who must make decisions, direct the orderly progress 
of the development program, and intelligently plan and 
control the work of the department. Cost records are ad- 
visable so that those responsible for the functioning of the 
research department can control the expenditures and de- 
termine the amount of funds available for subsequent opera- 
tions. Thus, the financial executives can assist in the direction 
of the research and development division towards the ulti- 
mate goal of all concerned: progress for the organization, 
growth, and additional profits. 

Operations research will help the business executive of 
the future make decisions more intelligently in various con- 
trol areas, but the decisions will always remain to be made. 
The possibility of removing all subjective and qualitative 
factors must be deemed at the present time to be more a 
hope than a real possibility, and the construction of com- 
pletely consistent and logical goals, while a reasonable ob- 
jective in decision making, may be unattainable. The balanc- 
ing of the responsibilities to society, consumers, owners, and 
employees will therefore still be the fundamental task of the 

The executive must continue to maintain control over 
operations. It has been suggested that the relationship of 
operating profit to gross assets is the best measure of operat- 


ing efficiency. Where control of cost cannot be readily meas- 
ured in terms of dollars, as in the case of broad investigative 
and research programs in manufacturing, engineering and 
sales, a project control may be used. 

Nothing paralyzes action on the part of executives like 
excessive analysis of figures. A good general rule is to carry 
the analysis of breakeven only far enough to give direction 
and stimulus to effective action on the part of the admin- 
istrator concerned. Only a few charts should be used to 
enable the top administrative officials to maintain a perspec- 
tive view of the enterprise and to direct and encourage ef- 
fective action on the part of managers down the line. In fact, 
almost anything an executive wants to know about the 
operating economics of the business enterprise can be learned 
by the use of breakeven point controls. 

The production process and the administrator have moved 
into such theoretical areas as the learning curve. The basic 
theory of the learning curve is simple: a worker learns as he 
works, and the more often he repeats an operation, the more 
efficient he becomes, so that the direct labor input per unit 
declines. What was not known until a decade or so ago is 
that the rate of improvement is regular enough to be pre- 
dictable. There is every indication that the learning curve 
offers a practicable answer to the needs of thousands of 
manufacturing companies for fairly accurate forecasts of 
direct labor requirements and productivity, but it is still a 
new device in a more or less "experimental" stage. 

On the marketing side the function extends far beyond 
sales. It entails cooperation and coordination with other 
parts of the organization, specifically production and finance. 
The marketing administrator must take into consideration 
the other functional areas so as to produce the kind of sales 
volume and dynamics that will give the company the great- 
est net income as well as the greatest potential growth. Cost 
accounting, properly interpreted, is essential to administra- 
tors for guidance in avoiding unprofitable sales and seeking 
to make sales which produce optimum profits. Such things 
as the allocation of fixed cost by territory should be avoided, 
because they could produce unrealiable or misleading fig- 
ures. However, marketing decisions based on figures alone 
would be ill-advised. 

The control of quality is a most important factor in the 


manufacture of material products; similarly, quality control 
is important with respect to the human beings who are en- 
gaged in the various activities involved in the manufacturing 
process. A major factor which should be kept uppermost 
in mind by those engaged in any phase of personnel quality 
control or by those contemplating the initiating of any of 
the techniques involved is that they are dealing with human 
beings rather than with machines or other inanimate objects. 
If a company whose top administrators are human-relations 
minded develops a competent staff of personnel technicians 
and gives them the authority commensurable with their re- 
sponsibility, the results in the form of improved quality of 
personnel should be evident in a relatively short time. 

The business firm needs a more adequate measure of or- 
ganizational performance than it is now getting. Progress in 
the social sciences now makes these measurements possible. 
Thus, top administrative officers need measures that provide 
them with data to fill the current serious gap in the informa- 
tion coming to them and to their organization. 




A. F. D. Campbell* 

Mr. Campbell emphasizes the importance of cash forecasting in this 
article and describes the steps involved in making such a forecast. 

Modern business has found it more necessary than ever before to take 
a watchful, searching and co-ordinated approach to the future. Forward 
planning has become one of the most serious preoccupations of top man- 

A fundamental element in forward planning is the establishment of a 
continuous balance between the cash position and the other aspects of the 
plan. Otherwise, of course, vital parts of it may be encumbered or brought 
to a complete halt through lack of liquid funds; or conversely, accumula- 
tions of cash may be immobilized, with consequent loss of potential 

Normally, the whole forward plan, of which the cash forecast is an in- 
tegral part, can be set forth in fairly realistic terms and in meaningful 
detail for several months ahead. Thereafter, the plan becomes of necessity 
more generalized and may be projected from three to five or more years. 

*From The Canadian Chartered Accountant, LXX, 4 (1957). 3 9" 12 - Reprinted 
by permission of the Canadian Institute of Chartered Accountants. 


Cash Forecasting i y i 

The short-term plan, being measurable against recent actual performance 
and current trends, will bear on those relatively tangible problems already 
facing the business and demanding immediate attention. The long-range 
plan is subject to such broad influences as national economic trends, the 
effect of technological and competitive progress on customer demand, and 
the development of new products and new markets. These forces direct 
management into continual reappraisal of manufacturing facilities, meth- 
ods of procuring materials, labor supply and of the future cost-selling 
price relationships. In all of this work, the cash forecast plays an im- 
portant role because it covers an area which is highly sensitive to changes 
of all kinds. 


The normal starting point of the cash forecast is the determination of 
the gross sales figures upon which the production, purchasing, facility and 
most other aspects of the forecast depend. It is therefore desirable that 
the sales projection be expressed in the greatest degree of detail for which 
there is logical support. Accuracy in sales forecasting is essential, and it is 
almost needless to remark that inaccuracy will render the cash forecast 
quite useless, if not harmful. 

The accurate determination of sales requires a thorough survey of the 
market available to the entire industry for products of a similar or allied 
nature. It is then necessary to contemplate the share of this market which 
is feasibly obtainable by the particular business, both on the short and 
long term. Here consideration must be given to the competitive status of 
its products in price, quality, reputation, and all other aspects bearing on 
customer acceptance. The development of such a sales potential will 
generate a number of questions which are worthy of careful study and 
management decision. To what extent does the sales potential match the 
existing manufacturing facilities, marketing organization and manpower 
supply of the business? Of the available market, what selection of products 
is best calculated to maximize sales of the profitable lines and still provide 
assurance of adequate volume to cover fixed charges? To what extent are 
the risks involved in expansion of facilities justified by the return which 
the attainable added volume will yield? If an investment is made in addi- 
tional plant, how long will it take to recover the cash so expended through 
the marginal cash revenue arising from the added volume? These or simi- 
lar questions need to be answered in conjunction with the proper de- 
termination of projected sales. 

For the purpose of the cash forecast it is, of course, necessary to plot 
the sales projections in terms of gross cash receipts. This procedure may 
be quite complicated depending on the nature of the business. Frequently, 

j - 2 Cash and the Administrator 

certain products or portions of certain product lines will be sold to cus- 
tomers whose payment habits vary from normal credit terms. Also, the 
payment behavior of the same customers may vary on a seasonal pattern 
throughout a year. Where the company is engaged in export operations, 
the cash forecast must make allowance for shipping delays and foreign ex- 
change transactions. Since the cash forecast deals in gross revenue, it is 
necessary to include in the sales revenue such additional items as freight 
and taxes, where it is the practice for the vendor to pay and may differ 
considerably among types of products and the geographical destination 
points of delivery. These are but a few examples to illustrate the factors 
which are encountered in the accurate forecasting of sales revenue. Once 
a business has gained experience in this type of close estimating, it is 
usually possible to develop a number of mathematical formulae, the use 
of which not only speeds preparatory work but also ensures that no 
factors are overlooked. 


After the sales projection, the next most important item in forward 
planning is usually the estimate of purchased materials. It is an area where 
the planning techniques can contribute tangibly to improved results. 

The sales projection provides a guide so that material requirements can 
be plotted into the future. With advance knowledge, the procurement of 
material purchases can be explored more thoroughly, and full advantage 
can be taken of competitive situations as well as seasonal fluctuations in a 
variety of commodity markets. Also, negotiations can be discussed in 
terms of larger quantities at more favourable prices. Therefore, while the 
buying will be based on the sales projection, there may be a considerable 
time lag between receipt of cash and the demands for cash disbursements 
created by the purchasing program. Advantageous purchasing arrange- 
ments may frequently cause a heavy drain on cash funds during a period 
when sales revenue is at a low point. Conversely, cash accumulations from 
sales may pile up in a later period when the business is realizing on an in- 
ventory bought in prior periods. 

Hence the cash forecast has a most beneficial influence in the forward 
planning of the purchasing program. Properly used, it enables management 
to foresee the need and test the availability of short-term loans or other 
sources of funds. It also makes it possible to calculate the duration of loans 
and thereby to negotiate the most suitable and favorable financing ar- 
rangements. It transpires that the purchasing program must be modified; 
this fact is made clear at a sufficiently early stage to allow the adjustment 
to be made at the least cost to the business. 

Cash Forecasting 173 


Of equal and often greater importance than the purchasing program 
are the financial implications of the labor bill. Modern techniques of 
forward planning are designed to enable a manufacturing company to 
make the best possible working arrangements for its labor force. While 
any production program must be kept sufficiently fluid to adapt to unfore- 
seen market conditions, the basic concept will be to maintain a steady flow 
of work in regular hours. Apart from the highly desirable results forth- 
coming from a congenial working atmosphere created by a level flow of 
work, important savings are obtainable when production is turned out by 
experienced hands and when overtime, training and other penalty costs 
are kept to a minimum. For this reason, the cash payout for wages may 
follow a different trend from cash receipts, although both factors are, in 
the long run, based on the sales projection. Here again the cash forecast 
fulfils a vital function in providing advance warning of any pressure on 
the cash resources which might arise from this situation. 

The preparation of the payroll section of the cash forecast is worthy of 
considerable research. The amounts involved are usually large and care 
must be taken to achieve a high degree of accuracy. Since the payroll takes 
time to prepare, the net wages will usually become payable several days 
after the end of the workweek or period. This time lag may cause dis- 
proportionately high or low cash requirements in certain months. The 
payment pattern of the various payroll deductions and fringe benefits also 
requires careful study. For example, the funding agreement of pension 
plans may be sufficiently flexible that the business can make payments at 
times in the year best suited to its cash position. 


Observations made on the cash forecasting of sales, purchased materials 
and wage payrolls apply in varying degrees to the many other items 
which are involved in the forward planning of business operations. Busi- 
ness practice has caused many expenses to be incurred and paid for at times 
in the year which are not logically co-ordinated with either the needs of 
the business or its cash position and frequently without any benefits in the 
form of reduced prices. Sometimes various supplies are carried in stock 
which are available at the same price on an immediate delivery basis. Per- 
haps the premiums of insurance policies fall due at times which are incon- 
venient to the cash position of the company. Stationery, office and many 
plant supplies may be bought in quantity in, say, January because it has 
become a habit to let the stock run low at the end of the year, regardless 

! -. Cash and the Administrator 

of the fact that January might be an awkward month for meeting cash 
payments. Ideally, the manner in which the expense budget is incurred 
and paid for should be arranged, as far as possible, to suit the trends of 
the cash position. 


When the receipts and disbursements arising from the trading activities 
of the business have been plotted on the cash forecast, attention can be 
turned to the demands which will be imposed on the cash position by the 
capital expenditure estimates. The accelerated pace of modern business cre- 
ates high obsolescence not only of tools, machinery, and buildings but also 
of the geographical suitability of entire manufacturing plants. In the past 
ten years, the size and frequency of these expenditures has become a mat- 
ter of major concern to most aggressive companies. While the operating 
budgets will reflect consequent increases in the depreciation and other 
amortization costs, the charges are nevertheless spread in the budget over 
a term of several fiscal periods. The cash forecast, on the other hand, must 
make provision for the entire outlays involved in new manufacturing 
facilities on, or close to, the dates of their acquisition. The function of the 
cash forecast is most important in equating this type of anticipated cash 
burden with future sources of funds. 

The rise in industrial expansion generally has created a wide variety 
of competitive sources and methods of long term financing. Business can 
now come closer than ever to a financing plan tailored to its individual 
needs, although it may find the money market more selective. 



Grover E. Edwards* 

Mr. Edwards concisely outlines the factors entering into a cash budget. 
The illustrations and the discussion pertain to an actual operation. 

After establishing the sales volume and after the operating budget has 
been approved, many companies have found it necessary to prepare a cash 

*From National Association of Accountants Bulletin, XXXIX, 3 (1957)1 6 7~73- 
Reprinted by permission of the National Association of Accountants. 

Structure of the Cash Budget 1 7 5 

budget. Why? Because, in today's business, it is necessary to know in 
advance by month, yes, even years, how much cash will be required for 
plant expansion, operating costs, and for increased receivables and inven- 
tories resulting from expected increases in sales volume. 

In our company the sales policy committee forecasts the expected sales 
volume for the coming years and the accounting department establishes 
the related operating budgets just as is done by many other companies. 
However, we go somewhat further and establish a budget for a sales 
volume that is somewhat higher and another for a sales volume that is 
somewhat lower than the original sales forecast. We also prepare cash 
budgets based on all three of the operating budgets. The two extra budgets 
are primarily a safety factor. Our cash budgets may show a drop in profits 
and a decrease in cash to a dangerously low point and we may need to 
obtain loans in advance of normal requirements to be on the safe side. On 
the other hand, the cash budgets may show that an increase in profits will 
cause an excess amount of cash and we will then be prepared to invest this 
cash or to pay off any loans that are no longer needed. 

In order to prepare the cash budgets, we made certain assumptions in 
addition to the ones already made for the operating budgets, such as: 

1. Minimum cash balance to be 4. Inventory increases and de- 
maintained, creases. 

2. Cash collections. 5. Plant expansion programs. 

3. Payment dates. 6. Depreciation policies. 


The minimum cash balance to be maintained must be decided upon be- 
fore preparation of the cash budget can be started. Although the cash 
budget, when completed, will enable us to prepare for the large anticipated 
cash expenditures, it is equally as important to provide an adequate cash 
balance for the day-to-day fluctuations resulting from normal plant opera- 
tions. Cash collections, payrolls and other expenditures are usually not 
spread evenly throughout the budget period. Cash collections may be de- 
layed for some reason or an unexpected expenditure of a considerable 
amount of cash may be necessary. Consideration must also be given to 
the minimum balances required by banks or other loan agencies, to main- 
tain credit for future loan requirements. 

In order to establish the minimum cash balance we needed, a study was 
made to determine what our cash balances had been during past periods. 
Based on this information, we decided on the minimum cash balance for 
the budget periods. Where two or more companies are to be consolidated 
for budget purposes, a minimum balance is established for each. Often 

1 76 Cash and the Administrator 

the movement of cash between companies is restricted by income tax con- 
sideration, by inter-company dividend policies, or other reasons. These 
must be taken into account. 

exhibit 1 

Cash Budget — Assumptions 

1. Cash balance to be maintained (minimum) 

2. Accounts receivable— 2/3 of previous month's net sales 

3. Notes receivable will increase $100,000 per quarter 

4. Inventory: Increased $800,000 11/30/56 and 2/28/57 

Decreased $300,000 5/31/57 and 8/31/57 
Extra inventory required due to addition of another plant 

$300,000 5/31/57 plus $500,000 8/31/57 

5. Plant Expansion program: 


Machine tools 
Bldgs. & other equip. 


ist Qtr. End. 1 1/30/56 
2nd Qtr. End. 2/28/57 
3rd Qtr. End. 5/31/57 
4th Qtr. End. 8/31/57 


$100,000 $ 
350,000 30,000 
400,000 50,000 
700,000 40,000 


Replacement Program 

ist Qtr. End. 11/30/56 
2nd Qtr. End. 2/28/57 


3rd Qtr. End. 5/31/57 
4th Qtr. End. 8/31/57 


(Depreciation will offset additions) 

6. Cash value of officer's life ins. increase $3,000 per qtr. 
Prepaid pension plan payments 

ist Qtr. Due 9/30/56 $50,000 3rd Qtr. Due 5/31/57 $10,000 

10/31/56 10,000 4th Qtr. Due 6/30/57 65,000 

2nd Qtr. Due 12/31/56 5,000 7/31/57 11,000 

(Others to remain approximately the same as beginning of quarter.) 

7. Accounts payable bal— 12/2 per cent of next qtr's. expected volume. 

8. Accrued liabilities: 

Accrued commissions— 4.1% of previous sales volume 

Cash profit sharing— 10% of previous quarter's net profit 

Profit bonus— 15.9% of net profit, payable in November 

Profit sharing savings— 10% of average monthly net profit for previous 

Accrued payroll— Proportionate to sales volume for following qtr. 
Accrued taxes— Sales taxes proportionate to vol. (.11%) 
Property taxes— $60,000 for 1957, payable 60% in May, 40% in Aug. 
Social security— same as 1955 except for add'l. employees 
Donations res.— 5% of net profit before taxes, less $5,000 per month paid 
All others unchanged from 8/31/56 

9. Federal income tax— Additions per profit schedule 

Structure of the Cash Budget 1 77 

Payments- 1 95 5-56 Payments-1956-57 

10% already paid 10% of est. tax due 5/15/57 

50% of bal. due n/15/56 $812,000 10% of est. tax due 8/15/57 

Balance— due 2/15/57 812,000 

10. Deferred income on installment notes was 36.6% of total notes receivable 
on 8/31/56. This percentage has been used for each quarter. 

11. Repayment of long-term debt, to maintain the $1,000,000 cash balance. 

12. Net worth and capital stock unchanged, surplus increased by net profit 
after taxes, cash dividends (present plan) $80,000 quarterly on Jan. 1, 
Apr. 1, July 1, and Oct. 1. 

13. Net sales— Annual Vol. $36,000,000 Net profit— $2,700,000 
1st Qtr. 20% 7,200,000 1st Qtr. ( 7 ,% f vo \ 54°> 000 
2 ndQtr. 20% 7,200,000 2nd Qtr. ^fter taxes' 54 ' 000 
3rd Qtr. 30% IO ,8oo,ooo 3rd Qtr. ^ b 810,000 
4th Qtr. 30% 10,800,000 4th Qtr. 810,000 
Note— For this example only one year of operations and expansion 

program are shown, though normally a budget should be 
prepared for a longer period especially when an expansion 
program is undertaken. 

When our cash balances decrease below the established minimum, we 
assume, for budget purposes, that additional funds will be provided by 
long- or short-term borrowings, whichever may be required. When cash 
balances increase above the required minimums, the excess cash is applied 
in the budget to reduce outstanding loans and, after these are paid off, is 
shown as invested in U.S. Government bonds. Actually the terms of a 
loan will govern repayments but, for budget purposes, payment of the 
loans is shown as described above, because the terms are not known when 
the budget is prepared. 

How soon after a shipment is made can we collect the cash— 10, 20, 30 
or more days? Will sales be made on conditional sales contracts under 
which payments are extended over a longer period of time, possibly up 
to several years? To provide us with information in this area, we pre- 
pared an analysis of previous monthly accounts receivable balances com- 
pared with the monthly sales volume. This comparison showed that the 
accounts receivable balances averaged 62 per cent of the previous month's 
sales volume. The study also brought out the fact that our collections 
average around 20 days from the date of shipment to the date of collec- 
tion. For cash budget purposes, we decided that the accounts receivable 
balances should be 66 per cent of monthly sales, to allow for possible ex- 
tension of credit terms. 

Conditional sales contracts, with payments being made on an installment 
basis or some other basis, will increase notes receivables balances and re- 
duce cash collections. Such sales are often necessary to increase, and some- 
times just to maintain, sales volume. Our notes receivable usually increase 

178 Cash and the Administrator 


Cash Budget - 

- Balance Sheet 

(Thousands of dollars) 

Sept. /, 

Nov. 50, 

Feb. 28, 

May 5/, 

Aug. 31, 


Cash and Gov't Bonds 
Notes Receivable 
Accounts Receivable- 
less reserve 











Total Current 




1 4,43 5 


Plant and Equipment— 
at cost 

Less Depreciation 
Other Assets 









Total Assets 



Accounts Payable 
Accrued Liabilities 
Federal Income Tax 











Total Current 
Deferred Income 
*New Loans 
Reserve for 







4,34 2 




Net Worth 

Capital Stock 

Less Cash Dividends 




( 80) 




( 160) 




( 240) 




( 320) 

Total Liabilities 
and Net Worth 



* New Loans is balancing figure to maintain cash at minimum balance. 

during the winter season due to promotion of three basic credit plans, gen- 
erally referred to as floor plans, longer terms and tonnage payments. They 
are as follows: 

Floor Plans— We make shipments of machines to our distributors dur- 
ing slack seasons, requiring payment to be made either when a sale 
is consummated or three to four months from date of shipment on an 
installment plan basis, whichever is the earlier. Discounts are allowed 

Structure of the Cash Budget 1 79 

if total payments are made on or before due dates of the installments 
and interest is charged if not paid by due dates. 
Longer Terms— We increase the length of time before a note is due, 

either on a single-payment or on an installment plan. 
Tonnage Payments-Under this plan, a certain dollar amount is payable 

based on each ton of material handled. 
Cash collections include not only collections from normal sales but the 
collections of royalties, engineering fees and dividends from affiliated or 
subsidiary companies, interest on notes receivable, and other miscellaneous 
income items. In addition, the cash made available through long- and 
short-term loans must be included. 


The payment dates of suppliers invoices, payrolls, profit-sharing plans, 
pension plans, plant expansion programs and federal income tax should be 
scheduled. We consider it generally desirable if cash is available to pay on 
a 10-day basis, especially when discounts are involved. For our cash 
budget, we consider all suppliers' invoices to be on a 10-day basis and fig- 
ure payments for a given month to be one third of the previous month's 
purchases and two thirds of the current month's purchases. All of our pay- 
rolls are considered to be distributed equally throughout the budget 
periods, even though paying our factory payroll every two weeks may 
cause them to fall unevenly in the budget periods. We feel that the varia- 
tion which results is not significant. Our pension plan premium payment 
is a substantial figure and the date for this payment must be scheduled. 
Federal income tax payment dates must be scheduled, along with pay- 
ments for contractor's invoices for the work completed in an expansion 

In a seasonal business such as ours, it is necessary to plan for inventory 
increases during certain months of the year and on decreases during other 
months. During the period of low shipments, from October throughout 
February, management must decide on the level of production which is 
to be maintained. The desire to maintain a stable working force with rela- 
tively steady hours of work and the degree of risk which management is 
willing to assume in order to be prepared for heavy demand during the 
construction season, are prime factors in this decision. Having decided on 
an optimum level of production and assuming the sales volume already 
estimated, it is quite simple to determine the inventory increase and later 
decrease. We do this by taking the difference between production vol- 
ume and shipment volume and then converting the sales value of the 


Cash and the Administrator 


Cash Budget — Application of Funds 

Quarter Ended — 

(Thousands of dollars) 

Nov. 30, 

Feb. 28, 

May 31, 

Aug. 31, 





Funds Provided by: 

New Loans 

2 74 

Net profit before Tax 





Decrease in Accounts Receivable 


Decrease in Inventory— Plant I 





Decrease in Other Assets 

* 3°° 


Increase in Accounts Payable 



Increase in Accrued Liabilities 





Increase in deferred income 









Funds Applied to: 

Repayment of Loans 


Increase in Notes Receivable 





Increase in Accounts Receivable 


Increase in Inventory— Plant I 



Increase in Inventory— Plant II 



Increase in Other Assets 

Purchase of Plant & Equip.— Plant I 





Purchase of Plant & Equip —Plant II 





Less Depreciation 

( 50) 

( 50) 

( 50) 

( 50). 

Decrease in Accounts Payable 


Payment of Cash Profit sharing 





Payment of profit-sharing savings 





Payment of profit bonus 


Payment of local taxes 



Payment of accrued donations 


l 5 



Income Taxes 














Excess Funds Provided (Applied) 

( 461) 

( 139) 



Cash— Beginning of Period 





Cash— End of Period 


( 461) 


I , I 35 


Increase (decrease) in cash position 

( 139) 


difference to cost through use of an average gross profit percentage. In- 
ventory requirements for new plants must be determined and allowances 
for them included in the cash budget. 

Structure of the Cash Budget 1 8 1 

Plant expansion programs usually run for a period from one year to 
several years. This type of expenditure requires a great deal of planning, 
not only for the program itself but for the cash requirements. The timing 
of construction of buildings and the delivery dates of machinery must be 
carefully scheduled so that, when the invoices are submitted from the con- 
tractors or suppliers the cash will be on hand. To assist in the prepa- 
ration of our budget, we plotted, by fiscal year quarters, the proposed 
scheduling of payments for the expansion program. 

The method of depreciation used for tax purposes is important in a cash 
budget. The faster write-ofT allowed by the government under the new 
declining-balance method and the sum-of-the-digit-method will give a 
quicker return of working capital. We use the declining-balance method 
on buildings because no salvage value is required. On the other fixed assets, 
we have found from experience that salvage values are insignificant. This 
combination of methods gives us the quickest return of working capital 
under the presently allowed methods of depreciation. 


At the same time the balance sheet is being prepared for the cash budget, 
we also prepare an application-of-funds statement. We find that carrying 
through both statements simultaneously helps to spotlight changes and 
provides a check on the accuracy of the compilation of the figures. 


To further illustrate the prerequisites for preparing a cash budget, Ex- 
hibit i shows the various assumptions it was necessary for us to make. 
From these assumptions the Cash Budget Balance Sheet (Exhibit 2) and 
Application of Funds Statement (Exhibit 3) have been prepared. Although 
figures are hypothetical, they are typical for a business of our type. 

The length of the budget period depends on several factors, including 
the purpose the cash budget is to serve, the financial condition of the com- 
pany, and the opinion of management as to the practicability and accuracy 
of estimating. Should the cash balance be low, an estimate of cash receipts 
and disbursement may be necessary on a weekly, or even a daily basis. 
However, a firm with ample cash may develop a cash budget by quarters, 
by six-month periods, or by years, whichever is considered necessary. Each 
company will have to establish its own budget periods according to the 
desires of top management. 

1 82 Cash and the Administrator 


Maurice Moonitz* 

Professor Maurice Moonitz discusses the statement of sources and ap- 
plications of funds in some detail, including the bothersome definition of 
"funds." The article is essentially a complete exposition of this important 

In the conduct of an enterprise, management has two major financial 
tasks of importance to accountants, namely, ( 1 ) to operate the enterprise 
profitably; and (2) to finance the activities of the enterprise and to keep 
it solvent. In the case of nonprofit activities, the statement of the first task 
should be modified to indicate that management is charged with making 
the enterprise perform within its prescribed limitations, such as a budgeted 
amount of expenditures. 

With respect to the first task, accounting has done a good job. That is 
to say, the necessity of a formal report on the results of operations is 
widely recognized, and the numerous problems involved in its preparation 
have received close and earnest attention among all groups and at all levels. 
With respect to a report on the way management has discharged its second 
task, our performance is less satisfactory. It is true, of course, that both 
the flow of funds into and out of an enterprise, and the effects of its 
financing activities, are recorded in the books, but a formal statement or 
report is typically not prepared. 

The statement of the sources and the applications of funds is an attempt 
to report on the second task of management, to fill a gap usually left open 
in the typical published report. The statement is, therefore, a supplement 
or addition to the conventional battery of statements, not a substitute for 
them in whole or in part. Assertions, for example, that a funds statement 
is better or worse than an income statement are unfortunate. The two 
statements have different functions; both suffer from attempts to set them 
up as rivals. 

The need for a statement to report changes and movements not clearly 
reflected in the balance sheet and the income statement has long been felt. 
In this country, the attempt to construct such, a statement is usually dated 
from the publication in 19 15 of William Morse Cole's Accounts: Their 
Construction and Interpretation. In that book, Cole described his "where- 
got-where-gone" statement. But progress was slow. In 1929, Myron M. 

* From The Accounting Review, XXXI, 3 (1956), 375-85. Reprinted by permission 
of the publisher. 

Reporting on the Floiv of Funds 1 8 3 

Strain wrote in his Industrial Balance Sheets (page 132), "The statement 
of application of funds had best be described, as it is one of the most useful 
of accounting statements and deserves frequent use; but it may be dis- 
missed briefly, because it does not get it. This exhibit details the sources 
from which all the funds used during a fiscal period were derived, and 
describes the uses to which they were put. It is a striking and significant 
interpretation of the changes that have taken place in financial position 
between two periods." Hector R. Anton, in the October, 1954 issue of the 
Accounting Review has published a report, "Funds statement practices in 
the United States and Canada." Anton notes that, according to his survey, 
68 per cent of the companies involved used a funds statement in some way 
or other, but that only 19 per cent included such a statement in annual re- 
ports to stockholders. 

The two managerial tasks under discussion are, of course, related. It is 
sufficient for present purposes merely that they are not identical. As a 
matter of fact they tend to merge into a single problem or task over the 
entire life of an enterprise, or, as a practical approximation, over a substan- 
tial time period. That is to say, over the long pull, a profitable concern 
will also be a solvent concern, although the reverse proposition, namely, 
that a solvent concern will also be profitable is manifestly not true. Over 
a relatively short period of time, however, profitability and solvency are 
almost independent of each other, sometimes almost antagonistic goals. 
Numerous cases are at hand in which enterprises expanded rapidly and 
profitably, but with a tremendous strain on working capital in the form 
of overextended receivables, swollen inventories, top-heavy current debt, 
and a marked shortage of cash. Similarly, other cases exist of concerns 
which are unprofitable for several years on end, yet actually improve their 
debt-paying ability in the process of contraction. The apparent paradox 
of profitability and solvency moving in opposite directions is not new, nor 
is it real in the sense of persisting indefinitely. But in the short-run, the 
two attributes pose two distinct problems; it is helpful to prepare account- 
ing summations of them at frequent intervals. 

To follow the point just made as to the relationship between profitabil- 
ity and solvency, we comment on the functions performed by the balance 
sheet and the income statement. The initial balance sheet of a newly- 
formed concern is usually also a good statement of funds— it reflects 
among the assets the results of the applications of funds acquired from the 
sources listed among the liability and proprietary items. Since operations 
have not commenced, the problem of profit-measurement does not arise. 
As a practical approximation, even a balance sheet prepared a little later 
on will also serve as a statement of funds. For example, this would be true 
in the case of a company with an extended development period during 
which little or no revenues arise, followed by an operating period in 

1 84 Cash and the Administrator 

which additional development work was negligible. But at some relatively 
early point along the way after operations have begun, the balance sheet, 
standing alone, begins to lose its function as a funds statement. 

At the other extreme, an income statement for the whole life-span of an 
enterprise would serve as the backbone of a funds statement covering the 
same time-interval. It would not be complete, because, to the revenues 
(sources) and expenses (applications) detailed in the income statement 
we would have to add at least (a) the long-term borrowings and repay- 
ments, (b) the issues and redemptions of capital stock, and (c) the divi- 
dends declared. But the combination of the items just listed would pro- 
duce an eminently satisfactory funds statement, without the intervention 
of a balance sheet. Notice that even the depreciation charge becomes an 
application of funds in this statement, because over a long enough period 
of time the summation of depreciation charges approximates quite closely 
the actual investment in the corresponding assets. As in the case of the 
balance sheet, some practical approximations can be introduced. For ex- 
ample, an income statement covering 50 years, supplemented by data on 
long-term financing and on charges and credits to surplus, would undoubt- 
edly be acceptable as a funds statement. But as the period covered by the 
income statement was shortened, the need for adjustments or modifications 
of some kind would begin to be felt in order to convert the income data 
into funds data. 

As a consequence, neither the balance sheet nor the income statement 
taken alone will give the story of the financial flows for the relatively 
short periods of time (one year, five years, ten years) covered in reports 
submitted to stockholders or prepared for top-management review. But 
these considerations with respect to the two conventional statements as 
potential exhibits of the flow of funds are worthwhile for two reasons, 
( 1 ) they indicate why a separate statement of funds is ordinarily desirable 
or even necessary for a complete reporting, and (2) they indicate the 
limited but by no means rare cases in which one of the conventional state- 
ments will double in brass and serve the purpose quite adequately of a 
complete reporting on management's performance of both tasks. 


According to Anton, 1 "in essence, funds analysis is the study of the flow 
of funds into the business unit and the uses for which such funds flow out 
during the same given time period." External transactions, then, are in- 
volved; so-called internal transactions, or transfers, amortizations, and 
accruals, do not constitute part of the funds flow. This emphasis is both 

1 Hector R. Anton, A Critical Evaluation of Techniques of Analysis of the Flow 
of Business Funds. Unpublished Ph.D. dissertation, University of Minnesota, 1953. 

Reporting on the Flow of Funds 1 85 

proper and important; in fact it constitutes the first principle underlying 
the statement. All the examples of funds statements that I have seen that 
are internally consistent, logical, and useful in throwing light on the 
financial activities of a concern make this distinction in some form or 

To help visualize the problem, a classification of financial flows is at- 
tached as Exhibit "A." Basically sound, the classification is not put forward 
as being necessarily the best one that can be devised. The purpose of the 
classification at this point in the discussion is to make fairly concrete the 
kind of thing we are talking about. A good funds statement, then, would 
include some or all of the items included in this classification; it would ex- 
clude other kinds of items found in the accounts of an enterprise or in its 
conventional accounting statements. For those who like to play with 
permutations and combinations, this classification is material for your 
recreation. Just by way of illustration, let us take the conventional form 
of funds statement as it has appeared recently in some annual corporate 
reports, that is to say, a statement accounting for variation in net working 
capital. Such a statement would show explicitly, if at all material, items 
1, 2, 5, 6, and 7 under sources as well as under applications; these items 
might be shown "broad" or "netted." Items 4 and 4(a) under both cap- 
tions would be combined in a net source of funds from operations (i.e., 
profit before depreciation and other "nonfund" charges to operations); 
item 3 would probably not be shown at all but instead would be buried in 
net working capital itself. 

If it has escaped attention, one other characteristic of the classification 
should be stressed. Each item in the classification constitutes part of a 
"flow," a movement, and refers to the amount received, for example, from 
customers during a given time period. None of the items is a balance on 
hand at any point of time. The scheme is therefore incomplete because it 
does not tie into anything. But it can be made complete (in the logical 
sense) by relating it to a "funds balance" at beginning and end of period; 
this can be done in several ways. One form which appears satisfactory is 
appended as Exhibit "B." 

Before the formal, technical problems of the statement are discussed, a 
word of caution may be in order. It is easy to become overly enthusiastic 
about the funds statement, an enthusiasm not justified by the capabilities 
of the instrument. True, the funds statement does supply information not 
otherwise available in conventional statements, but remember that it "re- 
verses the accruals" and ignores "internal transactions." Therefore, in at 
least one respect it is a cruder device than an income statement or a bal- 
ance sheet. The "cash-profit" approach of some discussions of this prob- 
lem can easily be overdone, and raise more issues than it resolves. 

1 86 Cash and the Administrator 

The positive uses of the funds statement and the reasons for the recent 
upsurge in its popularity are interesting. For one thing, the recent in- 
flationary movement in this country, associated with a high level of busi- 
ness activity and high tax rates, has posed financing problems on a scale 
so large as to constitute really new problems to American business. A 
statement of source and application of funds becomes useful in explaining 
why a net profit of a million dollars is not identical with an increase in 
funds of the same amount, available to increase dividends or raise wages. 
For another thing, the rapid changes of prices in an inflation make com- 
parisons of income statements difficult; a flow of funds analysis may help 
by submitting a more elementary, less sophisticated type of statement in 
addition to the income calculation. Finally, a further use, widely em- 
ployed by economists, and one in which accountants ought to develop an 
interest, is to reveal "distributive shares" in the output of a concern or an 
industry— how much "take-home pay," for example, does labor actually 
get, how much do the suppliers get, how much to creditors, stockholders, 
government, etc. Properly handled, a statement of funds is better adapted 
to the dissemination of this type of information than the conventional in- 
come statement with its highly abstract, sophisticated cost allocations and 
estimates, and its completely different orientation. 


We now proceed to the more formal aspects of the topic. Foremost 
among the problems involved is that of a fairly precise definition of 
"funds." A definition is necessary not only to satisfy the niceties involved, 
but also to assist us in the preparation of the statement and in the resolu- 
tion of new or difficult problems. A definition of funds provides this as- 
sistance by supplying a framework for the whole project, giving us a 
beginning and an ending balance into which the fund flows must fit or be 
reconciled, and thereby leading to a second "principle" by which to de- 
cide whether or not to include a financial event. 

The conventional statement of funds, as it has appeared in published 
annual reports of the last ten or fifteen years, will serve as a starter. 
Simply on the basis of frequency of appearance, little doubt exists that 
"funds" in these reports are defined as identical with "net working capi- 
tal," that is to say, the difference between current assets and current liabil- 
ities. This definition has the virtues of simplicity and of reliance on other 
widely-used concepts, namely, current assets and current liabilities. These 
virtues we will stress at the moment; its defects will be revealed later. 

Certain consequences flow from the definition. First of all, it establishes 
the content of the statement of funds— the statement must explain the 

Reporting on the Flow of Funds 1 87 

change, increase or decrease, in net working capital during the period 
under review. Secondly, and perhaps of more importance, it provides the 
basis for consistent and logical answers to any question that may arise with 
respect to the inclusion or exclusion of a financial event, because any ex- 
ternal transaction that increased net working capital is, by definition, a 
"source" of funds, and any external transaction that decreased net work- 
ing capital is, for the same reason, an "application" of funds. 

For example, under this conventional definition of funds, a stock divi- 
dend should not be reflected in a "funds" statement because a stock divi- 
dend does not involve any net working capital account in either its debit 
or its credit aspect. By contrast, a cash dividend, when declared, is an ap- 
plication of funds because it results in a credit to dividends payable or to 
cash, either of which reduces net working capital, and is therefore a part 
of the funds flow. Similarly, a dividend in kind, if it is payable in some 
current asset, also reduces net working capital and belongs in the funds 
statement; if, however, it is payable in a noncurrent asset (for example, 
the shares of stock of a subsidiary corporation) no working capital ac- 
count is involved. As a consequence, that transaction is not logically a part 
of the funds flow. 

The definition adopted also helps resolve questions with respect to 
amounts to be reflected and their classification, as well as with respect to 
the type of event to be included. Suppose, for example, that a substantial 
portion of plant is sold for cash at a loss. Ordinarily, the financial effect of 
this transaction will be reflected in the appropriate property accounts, 
their related depreciation allowances, and in an account reflecting the loss 
on sale of property. It may also involve offsets to income taxes. Two 
points seem clear enough, (1) the event provided funds, and (2) the 
amount to be reflected in a funds statement is the amount by which net 
working capital increased, in this case, the amount of cash received on the 
sale. But this amount is no longer reflected in a single account; therefore 
the bits and pieces of the transaction, as they are distributed through the 
property accounts, the depreciation allowances, the loss account, and 
possibly the related tax effects, should all be combined into one figure to 
show the increase in net working capital, the "source" of funds. 

But the definition of funds as identical with working capital apparently 
possesses certain disadvantages. Symptomatic of these is the recent experi- 
mentation with "cash flow" statements in published annual reports. For 
example, United States Steel and American Phenolic have presented two 
funds statements, one of conventional type, consisting of a schedule of 
changes in working capital during the year, and the other purportedly 
showing the flow of cash in and out of the business. These two state- 

1 88 Cash and the Administrator 

ments, as published in the 1952 report of American Phenolic Corporation, 
are appended as Exhibit "C." 

A record of cash receipts and disbursements, with the receipts classified 
by origin (e.g., from customers, issues of capital stock, borrowings from 
banks, etc.) and the disbursements by object of expenditure (e.g., to sup- 
pliers of materials, employees, stockholders, bondholders, etc.), while 
useful in its own way and undoubtedly a form of "funds" statement prob- 
ably is overly-narrow in its orientation. A better balance would be 
achieved if funds were defined somewhere between the narrow extreme 
of cash and the broad extreme of working capital; a useful middle point 
is the concept of "net money assets available for disposition." Concretely, 
this concept consists of the sum of cash on hand and in banks, marketable 
securities held as secondary cash reserves, and current receivables, less the 
current liabilities that will be paid by quick assets in the near future. In 
brief, funds become identified with cash on hand plus cash in process of 
collection minus checks in process of being written. Where bank financing 
is important, as in the American Phenolic case, bank loans can be excluded 
from the category of "funds" and treated as a source or application. 

A comparison of the conventional definition of funds with the one pro- 
posed above will indicate that the only major difference between the two 
is the treatment of inventories. Under the conventional definition, in- 
ventories are included as a part of net working capital, the funds balance 
itself. In the proposed definition, inventories are treated as a source of 
funds, when they are sold to customers, and as an object for the applica- 
tion of funds when debts are incurred to move them from the materials 
stage through process and into finished goods. This latter treatment seems 
more in accord with the function of inventories in a going concern. Of 
course, when the "inventories" are in reality indistinguishable from re- 
ceivables, they should be classed with the purely financial items. The 
reference here is to the output of a gold or silver mine or the work done 
on a cost-plus-fixed-fee (CPFF) contract. 

Regardless of the definition of "funds" adopted, certain types of finan- 
cial events that ought to be included will be omitted, unless the first 
principle previously enunciated is invoked. This first principle states that 
we are dealing with external transactions. The type of transaction that is 
likely to be left out, if attention is focussed too narrowly on the definition 
of funds, is the barter deal or the deferred-payment transaction. Take the 
case of a building acquired for 10 per cent cash and 90 per cent first- 
mortgage bonds. Whether funds are defined narrowly as cash, or less 
narrowly as "net disposable money assets," or broadly as net working 
capital, only the 10 per cent down-payment results literally and directly 
in a decrease (application) of funds. Still, the whole event is important, 
and, under the first principle, should be reflected by showing an applica- 

Reporting on the Flow of Funds 1 89 

tion of the full amount of the purchase price of the building and a source 
of funds equal to the bonds accepted by the vendor. When the bonds are 
retired, the statement for that period should show an appropriate applica- 
tion of funds to retire long-term debt. 

A rationale in the form of a presumed hypothetical intermediate cash 
transaction is theoretically satisfactory. Under this explanation, the event 
is treated as though the bonds were issued for cash, and the cash used to 
buy the building. But this type of explanation leaves the way open for 
other hypothetical interpretations which may not be so acceptable, and 
furthermore, is unnecessary. It is better, with any definition of funds, to 
refer back to our first principle, namely, that we are dealing with relation- 
ships between the concern and the outside world, and include these barter 
deals and deferred-payment arrangements explicitly, rather than by the 
back-door of hypothetical intermediate transactions. 

Certain other characteristic problems arise in the preparation of a funds 
statement. Noncash gifts and subsidies, for example, have no impact on 
funds; yet the amount involved may be material and the reporting essen- 
tial to a full disclosure of the way in which the enterprise is being fi- 
nanced. These gifts and subsidies should be included as (a) funds pro- 
vided by the donor and (b) applied to the object received. Gifts or 
subsidies in the form of cash do increase funds, however defined, and will 
appear in a funds statement. 

An example of this type of problem comes to mind. A hospital, newly- 
formed, acquired equipment for cash, and then was reimbursed for its 
actual expenditures by a governmental agency. In this form, no difficulty 
arose in the preparation of a funds statement— the hospital had clearly 
applied funds to acquire equipment; the hospital had clearly received 
funds from an outside source when it was reimbursed for its earlier out- 
lays on equipment. But suppose the governmental agency had acquired 
the equipment itself and made the gift (subsidy) to the hospital in kind. 
No cash (or other fund account) would have been involved on the hos- 
pital's records. Yet it seems clear that the two forms lead to identical 
results; the substance should prevail in the preparation of a funds state- 
ment as in the case of any other accounting report. 

One additional observation may be useful. Notice that we are not con- 
cerned with the classification of a gift or a subsidy as an increase in capital 
or in earnings, or in neither. We are concerned solely with the fact that a 
financial event occurred involving the entity and an outsider; when we 
report that event, we have fulfilled the requirements of a funds statement. 
In the related but not identical problem of income measurement or the 
reporting of financial position, the question of the proper classification of 
a gift or subsidy as between capital and income will have to be faced. But 

1 90 Cash and the Administrator 

not in a funds-flow analysis. As a consequence, a funds statement will re- 
flect the event in identical fashion, regardless of the manner in which it 
was reflected in the records— records which are conventionally designed 
primarily to assist in the determination of periodic income. 

Another problem is the treatment of depreciation and other amortiza- 
tion. No extended analysis of this warhorse is necessary; instead a few 
observations will be made. First, the application of funds to the deprecia- 
ble or amortizable item is reported in the period of its acquisition. Second, 
depreciation itself is omitted from the funds statement because it is a cost 
or expense, properly recognized in the measurement of income, which 
does not require the application of funds, however defined, in the current 
period. Third, any attempt to show depreciation as a "source" of funds 
is awkward, unnecessary, misleading, and just plain wrong. The reference 
here is to the widespread practice of adding back depreciation to the net 
profit figure to get the amount of funds provided by operations. This is a 
worksheet adjustment, and does not belong in a formal statement. The 
figure we are after, and that we usually get by this adjustment, is the 
amount of funds provided by operations before deducting a nonfund 
item such as depreciation expense. Fourth, the depreciation adjustment 
may be incomplete in a manufacturing concern— a considerable amount 
of depreciation may be tied up in inventories, and ought to be reversed. 

Another problem is the tendency to want to reverse the entries for 
estimated uncollectible receivables. Except where funds are defined as 
cash, and the funds statement accordingly becomes a report of cash re- 
ceipts and disbursements, this type of reversal is not warranted. Current 
receivables are a part of funds; the allowance for bad debts is an attempt 
to reflect those receivables on a net collectible basis, and should therefore 
be left alone in a funds analysis. Perhaps the difficulty arises when the 
analyst recognizes quite correctly that the charge to income for bad debts 
is not an application of funds. But the proper treatment in this instance 
is not to reverse the entry as a nonfund adjustment, similar to depreciation. 
Rather the charge to income should be interpreted as a revenue-deduction 
item, a correction of an otherwise overstated revenue account. If the bad 
debts debit is so interpreted, and it is the correct interpretation, no diffi- 
culty on this score will be encountered in the preparation of a funds 

A loss on the conversion of any funds item constitutes an outflow of 
funds, a diminution in the "pool" of homogeneous elements; as a conse- 
quence the loss would usually be classified as an "application" of funds. 
For example, assume that marketable securities, held as a secondary cash 
reserve, and reflected in the books at their cost of $100,000 are sold this 
period for $95,000. Assume also that in this same period a theft of $5,000 

Reporting on the Flow of Funds 1 9 1 

cash takes place, without recovery of any sort. Each loss of $5,000 repre- 
sents an "outflow" of funds, and should be so reported, even though the 
events themselves were unplanned and undesirable. The related case of a 
gain on the conversion of a funds item is clearly an inflow of funds, classi- 
fied usually among the sources. 

The treatment of inventories as a source of funds or an object of their 
application has already been urged primarily on the basis that inventories, 
in the usual case, are too important to be buried in a net working capital 
figure and require substantial outlays to move out to customers. 2 The sales 
figure is of course identical with the funds provided by customers during 
the period. The application of additional costs to process the inventory in 
the current period can be calculated in total by a simple formula, namely, 
the cost of goods sold plus the difference between the ending and the be- 
ginning inventories. In the case of the so-called actual cost systems, this 
formula will always hold regardless of the method of inventory pricing 
employed, whether cost or market, first-in, first-out, last-in, first-out, or 
average cost. The formula yields a total figure; it will not give the break- 
down of the costs among labor, materials, supplies, etc. 

In the case of a standard cost system, the formula just given will also 
hold, provided the variances are closed out at the end of the period to in- 
ventories and to cost of goods sold. If the variances are instead carried 
direct to income, the formula, as it stands, will calculate funds applied at 
standard, which is presumably not satisfactory in a funds analysis. Conse- 
quently, the formula should be expanded to include "plus or minus the 
standard cost variances." 

As the last problem to be discussed, consider the situation when a previ- 
ously non-current item becomes current, without an actual transaction 
with an outsider. Specifically, consider the case of the current portion of 
a serial bond issue. Each year a new series is detached from the long-term 
debt and placed among the current debts, indicating payment in the near 
future. In a funds statement, this amount is treated as an application for 
the same reason that a dividend declaration is so treated, namely, that 
payment in the normal course of events is automatic in the short-term. As 
a consequence, the pool of net disposable money-assets or of net working 
capital is diminished. In either case we have a clear case of an event giving 
rise to a decline in funds. 

2 Where the finished goods or merchandise is virtually as good as cash the con- 
ventional inclusion of inventories in the funds total is satisfactory. In addition to in- 
ventories of precious metals and costs tied up in CPFF contracts, inventories of com- 
modities for which a highly-organized spot and futures market exists would qualify 
for inclusion. The case for inclusion is especially strong if these types of inventories 
are stated at net realizable value instead of at cost, because cost does not, except by 
coincidence, measure the inflow of funds from the holding of highly-marketable in- 

1 92 Cash and the A dministrator 


In the process of preparing the funds statement, several methods are 
available. Vatter, of the University of Chicago, has proposed the deriva- 
tion of data from direct posting to T-accounts. 3 

Others have stressed the desirability of inserting a summary analysis of 
nonfund accounts in audit working papers, thereby also obtaining di- 
rectly the necessary data, as under Vatter's proposal, but without the in- 
tervention of actual accounts. This procedure is illustrated in Finney and 
Miller, Principles of Accounting: Intermediate, 4th Edition, and in Holmes 
and Meier, Intermediate Accounting, Revised Edition. The most widely- 
used method, however, is the process of adjusting changes derived from 
comparative balance sheets, as supplemented by an analysis of income and 
retained earnings. 

If this procedure is followed, certain technical problems arise which 
are similar regardless of the definition of funds employed. These may be 
summarized briefly as follows: 

1. Reverse the differences in account balances representing transac- 
tions not involving funds; (example: the depreciation entries) 

2. Reinstate any transactions involving funds that are suppressed in 
the usual accounting process; (example: sale of a non-current asset) 

3. Combine and reclassify the remaining items to bring the bits and 
pieces of funds data together. 

In the worksheet itself, the funds analysis should be quite detailed in 
order to insure that no important aspect of the financial flow has been 
omitted, overlooked, or underestimated. But in the statement itself, as in 
any financial statement that we prepare, similar items should be judiciously 
combined and grouped, important aspects played up, and minor, inconse- 
quential flows thrown into a "miscellaneous" or "all other" category. No 
one would disagree with the basic soundness of these commonsense rules 
of presentation of data. But a related problem lurks in the background on 
which there is no unanimity of opinion or of practice. The reference is to 
the extent to which similar sources and applications should be set off 
against each other. For example, all would agree that if X Company bor- 
rowed $1,000,000 from each of three different banks, we would meet all 
the niceties of disclosure and relevance if we reported a source of funds 
of $3,000,000 from bank borrowings; no one would insist that we ought 
to spell out the three separate borrowing operations. But suppose during 
the same period, X Company also paid off $1,500,000 of other bank loans. 
Should we now report a source of funds of $3,000,000 and an application 

3 Wm. J. Vatter, "A Direct Method for the Preparation of Funds Statements," 
Journal of Accountancy, June 1946, pp. 479-89; also see "Correspondence" in the 
September 1946 issue of the same Journal, pp. 256-57. 

Reporting on the Flow of Funds i o^ 

of funds of $1,500,000 or should we be content with the disclosure of a 
net source of funds of $1,500,000 from bank borrowings? 

Of more substance is the treatment of the funds flows generated by 
operations. Specifically, the reference is to the source of funds tapped by 
sales to customers and the applications of funds to wages, materials, etc. 
The reflection of a single source of funds from operations, calculated by 
adding nonfund charges to net profit, is found in most published analyses 
of changes in net working capital. But this practice may be omitting data 
on significant changes. Notice, for example, the difference in mode of 
treatment of the operating items in the American Phenolic data attached. 
In the statement of changes in working capital, funds provided by opera- 
tions are shown conventionally in two figures: net profit for the year, and 
provision for depreciation. In the statement of cash receipts and disburse- 
ments, however, the influence of operating items is reflected in one item, 
sales, under receipts, and in four items under disbursements (specifically, 
materials, supplies and services; salaries and wages; taxes; and interest). The 
recommendation here being urged is to set forth a funds statement more 
along the lines of American Phenolic's "cash flow" than along the lines of 
the same company's working capital analysis. Fundamentally, the point be- 
ing made is that we should guard against unwarranted inferences as to 
causal relationships, particularly when an application of funds is sub- 
tracted from an important source of funds, and the difference only then 
set forth in a formal statement. 

Part of the difficulty here stems from the fiction that the funds state- 
ment is an attempt to explain what became of the profit, and that ac- 
cordingly the tie-in must be with the net income figure. But even a casual 
examination of published funds statements will reveal that (a) they 
display sources and applications of funds beyond those connected with 
operations, and (b) they tie in with net working capital, or cash, or some 
other concept of funds, but not with net profit. No one can ever tell what 
became of net profit, a calculated magnitude, the difference between 
revenue and expense. We can tell a great deal however about the inflow 
of funds generated by sales, by borrowings, by issues of shares, and by 
other means, and the outflow of funds related to the services of employees, 
of suppliers, of lenders, of stockholders, etc. The influence of income- 
measurement is strong; it has obviously dictated the central position of the 
income statement. It has less obviously but nevertheless just as certainly 
dictated the form and content of the balance sheet. To judge by pub- 
lished statements, it has also influenced the form and content of the funds 
statement. But the proper function of a funds statement is not to tell us 
more about the income-generating and income-measuring processes, but 
rather to disclose data on the related but nevertheless distinct task of 
financial management of the enterprise. 

j g 4 Cash and the Administrator 



Funds are derived from 

i. Contributions of stockholders; 

2. Long-term loans, e.g., mortgages, bonds, equipment contracts; 

3. Short-term loans supplied primarily by commercial banks; 

4. Sales to customers. This class includes reduction of inventories; 

4(a). Government subventions not included in (4), above, such as subsidies to 
airlines, steamship companies, etc.; 

5. Disposal of non-current investments; 

6. Disposal of plant, property, and equipment; 

7. All other sources, e.g., gifts. 
Funds are used to 

1. Cover dividends and redeem shares of stock; 

2. Services and retire long-term debt; 

3. Service and pay short-term loans; 

4. Cover operating costs, such as labor, materials, supplies, etc. This class includes 
increase of inventories; 

4(a). Pay taxes not included in (4), above; 

5. Acquire non-current investments; 

6. Acquire plant, property, and equipment; 

7. Cover all other applications, e.g., loss by embezzlement. 



For Period froin to 

Funds, beginning of period xxx,xxx 

Sources tions 

Fund changes during the period: 

I. Net funds from operations (profit or loss as adjusted 

for non-fund items) -See Note xxxx 

II. Funds transactions with stockholders 

Dividends paid xxxx 

Investments , xxxxx 

III. Funds transactions with long-term creditors 

Sale of bonds xxxx 

Retirement of bonds : • • xxxxx 

IV. Funds transactions, involving plant and inventories, 


Plant, intangibles, and investments xxxx 

Decrease in inventories ; xxxxx 

T/otals xxxxx xxxxxx 

Net increase (decrease) 'in 'funds (See Schedule A) [Not reproduced] .. xxx,xxx 
Funds, end of period xxx,xxx 

Adapted from Hector Anton, op. cit., p. 107. 

Note: Anton concludes that, on balance, a reflection of net funds from operations fits m most 
closely with current practice and its apparent objectives. My own preference is for more detail, at 
least to the extent of revealing sales and the major operating costs. But in either case, the classifica- 
tion and form illustrated above will serve the purpose. 

Reporting on the Flow of Funds 1 95 



Statement of Changes in Working Capital 

Year Ended December 5/ 

i9S 2 l 95i 

Working Capital-Beginning of Period $ 2,904,385 I 3J4547g 

Funds Provided- 
Net profit for year $ 1,279,290 $ 941,868 

Provision for depreciation 54 : ,786 367 ,4 J l 

Proceeds from sale of fifteen year 4%% sinking fund notes 2,000,000 — 

Sundry, net 201,879 5Q,97i 

Total funds provided % 4,022,955 $ 1,360,250 

Funds Applied- 
Additions to plant and equipment $ 1,050,481 $ 1,673,326 

Provision for sinking fund including payment of long- 
term loans 969,049 232,950 

Dividends declared 380,532 320,016 

Increase in prepaid expenses 21,326 75,°49 

Total funds applied % 2,421,388 $ 2,301,341 

Net increase or decrease in working capital $ 1,601^567 $ 94 I ,°9 I 

Working Capital-End of Period $ 4,5Q5,952 $ 2,904,385 


Statement of Cash Receipts and Disbursements 

Year Ended December 31 

1952 1951 

Cash Balance-Beginning of Period $ 1,295,109 $ 808,424 

Sale of merchandise to customers $36,456,101 $24,355,836 

Bank loans, including $3,200,000 "V" loan — 3,950,000 

Sale of U.S. Government securities — 1,185,315 

Refund of prior year's federal income taxes and renegotia- 
tion - 45,261 

Long-term 4% % loan 2,000,000 — 

Sundry ioi,575 110,941 

Total receipts $38,557,676 $29,647,353 

Materials, supplies and services $21,229,937 $19,106,502 

Salaries and wages 8,312,128 6,572,904 

Taxes, including purchase of U.S. Treasury tax savings 

notes 4,195,700 1,205,563 

Plant and equipment 1,050,481 1,673,326 

Dividends 360,408 320,000 

Retirement of bank loans and "V" loans 1,450,000 — 

Debentures purchased, including deposits with Trustee 

for retirement of long-term loans 978,825 170,4! l 

Interest 254,733 103,962 

Total disbursements $37,832,212 $29,152,668 

Net increase in cash balance $ 725,464 $ 494,685 

Cash Balance-End of Period $ 2,020,573 $ 1,295,109 

!q6 Cash and the Administrator 


Bernard Whitney and Marion S. Israel* 

Experiments with reports and methods of assembling and presenting 
data are of considerable value in improving the understanding of often 
overlooked relationships and facts. The authors sketch the outline of an 
interesting model which combines facts of the balance sheet, operating 
statement, and application of funds. 

The time-honored accounting statement depicts a frozen slice of an 
enterprise, cut at an instant of time, usually many transactions in the past. 
It would be very useful to have a model of the financial structure of a 
business, which would not only show the dynamics of what is occurring, 
but also indicate why it is happening. It is most important in planning 
ahead to recognize the interrelations or flow of funds within a business 
that support or detract from one another. Excluding those few people 
who have developed highly intuitive financial insight, it is difficult for 
most managers to derive sound analytical judgments concerning the fi- 
nancial dynamics of a business unit by examining the balance sheet and 
profit and loss statement. There are many figures on separate statements 
that must be correlated, and the sources of change are not apparent. 

In this paper we do not propose to analyze whether traditional account- 
ing statements portray the true financial picture. Nor is there need to 
elaborate on the time and labor required for the compilation of modern 
accounting statements. With few exceptions, businessmen consider them 
excessive, although many in the accounting field disagree. An extension 
of the method suggested here has made possible the successful solution of 
complicated accounting problems taken from CPA examinations in less 
than one-third the usual time. 

Our analytical model for depicting the financial situation is a geo- 
metrical construction that combines three standard accounting presenta- 
tions: the Balance Sheet, Operating Statement, and Application of Funds. 
These are combined to represent quantitatively: (a) what has occurred 
in the recent past, (b) the dynamics of the current situation, (c) present 
trends, and (d) the readjustments of assets and liabilities required to ac- 
complish a given end. The model is intended to simplify the accounting 

* From Operations Research, VI, 4 (1958), 573-79. Reprinted by permission of 
the authors and the Operations Research Society of America. 

Financial Dynamics of a Business 1 97 

task and give management, after familiarization with its different form, a 
readily understandable and more complete concept of the financial inter- 
relations within the business. 

The illustration presented in this paper depicts on a single page in- 
formation which would ordinarily require nine pages for the three-month 
period. The accounting formulations employed have been in use many 
years. The method is basically simple, but the development of the model 


Balance Sheet, March 31, 1957 


Current assets 

Cash on hand $ 1,000 

Cash in bank 105,000 $106,000 

Accounts receivable $274,400 

Less provision for bad debts 12,400 262,000 

Inventories 97,000 

Total current assets $465,000 

Fixed assets 

Machinery and equipment $304,650 

Less depreciation allowance 955227 $209,423 

Furniture and fixtures $141,954 

Less depreciation allowance 39,377 102,577 

Total fixed assets 3 1 2,000 

Total assets $777,000 


Current liabilities 

Accounts payable $ 97,162 

Accrued wages 2,348 

Taxes payable 1 ,490 

Loans and notes payable 6,000 

Total current liabilities $107,000 

Net worth 

Capital $493,000 

Less drawings — $493,000 

Undistributed profit (loss) to date 177,000 

Total net worth 670,000 

Total liabilities and net worth $777> 000 

Cash and the Administrator 


Condensed Balance Sheets 

Jan. i Jan. 31 Feb. 28 Mar. 31 

Cash in banks $ 50,000 $ 76,000 $1 15,000 $106,000 

Accounts receivable 42,000 106,000 68,000 262,000 

Inventories 160,000 1 24,000 108,000 97,000 

Total current assets $252,000 $306,000 $291,000 $465,000 

Fixed assets 270,000 260,000 322,000 312,000 

Total assets $522,000 $566,000 $613,000 $777,000 

Accounts payable $ 29,000 $ 31,000 $ 47,000 $107,000 

Capital 493,ooo 493,ooo 493,ooo 493,ooo 

Plus profit 42,000 73,000 177,000 

$535,000 $566,000 $670,000 

Total liability and capital $522,000 $566,000 $613,000 $777,000 

Condensed Operating Statement 

January February March 

S a i es $210,000 $174,000 $390,000 

Material cost of sales 

Opening inventory 160,000 1 24,000 108,000 

Add: purchases 35,ooo 38,0°° 112,000 

Total $195,000 $162,000 $220,000 

Less /'closing inventory 124,000 108,000 97, 000 

Material cost $71,000 $54,000 $123,000 

Gross operating profit 


Other expenses 

Total expenses 

Net profit 


1 20,000 






153 ,000 



$ 89,000 

$ 31,000 


into a practical device is the product of considerable effort. The new 
element which has been added is the juxtaposition of the old, familiar 
figures so that their interrelations are made clear. 1 

Table I shows a slightly modified version of the Balance Sheet of an 
existing company. Table II is the corresponding Profit and Loss Statement. 
With traditional accounting methods, each of these requires three pages, 
one for each month. Figure 1 presents a single conceptualization or model 

1 In the development of this model, we have borrowed extensively from papers by 
Miss Helen Sunley. 

Financial Dynamics of a Business i 99 

which includes on one page not only the equivalent of the traditional 
Balance Sheet, Operating Statement, and Application of Funds, but addi- 
tional information showing what has been occurring in the interrelations 
between these different viewpoints. The difference sectors of the circle 
together represent a complete financial picture of the business: Accounts 
Receivable, Cash, Accounts Payable, Other Assets, and Proprietorship. 
The concentric rings or annuli depict the data for three consecutive 
months in this example. 

The "Payments" made by the business are in their proper relation when 
placed on the radial line which separates Cash and Accounts Payable. 
Similarly, "Purchases" are correctly located on the radius between Ac- 
counts Payable and Other Assets, "Usage" between Other Assets and Pro- 
prietorship, "Sales" between Proprietorship and Accounts Receivable, and 
"Collections" between Accounts Receivable and Cash. The logic of this 
arrangement is confirmed if we adopt the following method of showing 
the beginning and ending balances for each month and the differences be- 
tween them: the balances in each account are shown on the appropriate 
"date circle," in the center of the appropriate sector. Thus, for example, 
the balance in Accounts Receivable on January 1 was 42 ($42,000). On 
February 1, it was 106, and on March 1, 68. 

Looking outward from the center of the model, on the lefthand or 
counterclockwise side of each sector representing a particular account in 
a given month is the addition to that account which occurred during that 
month. In the case of Accounts Payable for January, the addition was 
120. At the right side of this same sector is shown the total deduction from 
the account, in January, 122. The net difference or increment between 
this income and outgo is minus 2, as shown in the delta (triangle) in the 
center of this circular segment. Since the balance of Accounts Payable at 
the end of January is the beginning balance for February, it is shown as 
minus 31 located appropriately on the concentric circle marked Feb. 1 
representing the transition of January to February. This quantity, minus 
31, is the negative increment of income-outgo (minus 2, referred to 
above) added to the January 1 Accounts Payable balance, shown as 
minus 29. 

As the circle is "rotated" or examined from different viewpoints along 
its periphery, we see that the additions to Accounts Payable along the 
Payments radial (120, 173, 205) are also the Cash outgo (at the right or 
clockwise side of the Cash sector). The collections, which as additions to 
the Cash account are shown at the left of this sector along the Collections 
radial (146, 212, 196), are in turn the outgo for Accounts Receivable. 
Similarly, Sales, which are income for Accounts Receivable, are also the 
outgo for Proprietorship. 


Cash and the Administrator 



Balance sheet, operating statement, and application of funds, first 
quarter, January 1 — March 31 (thousands of dollars). 

Figure i 

In the Other Assets segment of Fig. i, a further breakdown is ex- 
emplified. The segment of the concentric circle representing this account 
for March has been subdivided into three parts: an outer portion for 
"fixed asset adjustment," one for "inventory adjustment," and another for 
very "liquid assets." The same conventions are maintained with income to 
the left facing outward from the center of the model, outgo to the right, 
and the increment or difference in the delta triangle between. 

Because the quantities it portrays are derived from accounting state- 
ments that must balance and because of its construction, there are certain 
sequences within the model that must balance out or total correctly. The 
quantities on any full circle must total zero algebraically (equivalent to 

Financial Dynamics of a Business 201 

the zero difference or balance between Other Assets, and Liabilities and 
Proprietorship, on the standard accounting Balance Sheet). Thus, for 
example, on February 1: +76, +106, —535, +384, and —31 =0. The 
sum of the quantities in the deltas located within the same concentric 
circle must also add agebraically to zero (equivalent to Profit and Loss). 
Examining the figures for the month of February in Fig. 1, we find (be- 
ginning with Cash): +39, —38, —31, +46, — 16 = o. Likewise, within 
any sector (Cash, for example), the beginning balance for any month (in- 
ner circle) must add to its delta to give the closing balance (outer circle). 
For February: +76, — |— 39 = +115. And finally, the income and outgo 
within any one annulus must balance with the incremental difference 
shown in the delta. In Proprietorship for the month of March: —390 (out- 
go or credit) +286 (income or debit) = — 104 (delta). 

These balancing sequences provide easy and automatic checks for the 
manager who is not thoroughly familiar with the traditional accounting 
statements expressed separately and usually on different sheets. Equally 
important, the form of the model emphasizes the necessary interrelations 
not only between the quantities themselves, but what they represent con- 
cerning the financial operations of the business. 

For example, the model shows clearly that: profit = (increase in assets) 
-|- (decrease in liabilities). In March, for example, the profit of 104 ($104,- 
000) is composed of: Accounts Receivable, +194; Cash, —9; Accounts 
Payable, —60; Other Assets, —21 (inventory adjustment, —11; fixed asset 
adjustment or depreciation, —10; and o change in liquid assets). The 
March profit of 104 is also readily seen to represent the difference between 
Sales of 390 and Usage of 286. 

This form of analytical report can be prepared in a few hours from the 
usual accounting journals or from a specially designed summary journal. 
This eliminates waiting for the posting of the general ledger and speeds 
up crucial financial information by days or even weeks. The model is 
easily revised weekly, daily, or irregularly when needed, in contrast to 
the inflexible monthly schedule normally required for all practical pur- 
poses by traditional accounting methods. 

The model is also useful for planning purposes. Referring again to Fig. 1 
the executive can project the situation for the forthcoming month 
(in this case, April) in accordance with certain assumptions he wishes to 
make. Let us suppose he would like to double Sales again. From previous 
experience, he can estimate quite closely what his Accounts Payable and 
Accounts Receivable will be at this assumed Sales volume. Likewise, he 
can project probable Collections and Usage. With these data, Cash re- 
quirements are calculated readily and he may discover these are beyond 
his present means. Failure to recognize the cash requirements of ex- 
panded operations is in fact a common error in planning frequently made 

202 Cash and the Administrator 

by smaller enterprises. Various ratios, limits for accounts payable, and 
purchase requirements can be determined in the same way. 

In performing this projection into the future, the executive is required 
by the form of the model to identify and make certain estimates, to com- 
bine his assumptions and calculations based on empirical ratios in their 
proper interrelation so that they balance, and he cannot omit any one of 
the necessary financial components. Once he is familiar with these inter- 
relations, he may prefer that the circular model be 'unrolled' and pre- 
sented in tabular form. 

The concept of differentials expressed by the quantities within the 
deltas in the model can be incorporated into the regular accounting sys- 
tem by recording a "third entry" of the differentials reflecting the changes 
between past and current balances in the accounting ledger. Companies 
which have adopted this method of differential accounting obtain finan- 
cial reports that depict the flow of goods and money through the enter- 
prise and show the causation of changes in assets, liabilities, working capi- 
tal, and profit components. An accounting matrix composed of all the 
elements of the financial process is used to develop optimum strategies, by 
varying anticipated changes in sales volume, costs, collections, or other 
components subject to manipulation by management. 




Herbert J. Richmond* 

Mr. Herbert J. Richmond lists and discusses various factors and con- 
siderations which are important in establishing inventory and related 
production policies. One of the important points is the analysis and con- 
trol of inventory components on a value basis. 

Many business systems and policies which are sufficient for current 
conditions may prove inadequate in the future because of the increased 
pressures for automation and the guaranteed annual wage. The constant 
drive for lower sales prices and enlarged productive facilities will also put 
a strain on those systems and policies which now seem satisfactory. 

Are your company's policies and procedures for inventory control and 
production planning established to meet future requirements and to give 
maximum profits, stabilized production and good relations with custom- 
ers, vendors and employes? 

What steps are being taken to insure that men and machines will be 
serviced on a continuous basis with a consistent supply of parts and 
materials and with a minimum of paperwork? 

# From The Controller, XXIV, 4 (1956), 162-64, 182, 184-85, 196. Reprinted by per- 
mission of The Controller. 


204 Inventory and the Administrator 

Programs which will give affirmative answers to these questions may 
vary from company to company, depending upon the industry and the 
organization's policies, procedures and finances, but there are general 
points relating to inventory planning which can be considered in apprais- 
ing these matters. 


It is extremely important that the individual elements of the inventory 
be carefully analyzed since control and planning begin with the individual 
part and material. Each item in the inventory has its own characteristics 
which will determine its control. The annual usage of the item is extended 
by its unit cost to determine the annual expenditure for that part. The 
value basis is the best guide for classifying parts and for applying the de- 
gree of control necessary to recognize the dollar importance and to mini- 
mize inventory investment and the costs of control, such as paperwork, 
set-up, material handling, costing and issuance. 

In many companies the same policies, procedures and paperwork are 
applied to an item whose total yearly usage is $80 as is given to an item 
whose total annual usage is $80,000. This is true despite the fact that the 
numbers of items of high value in any inventory usually represent a small 
proportion of the total items, although their dollar proportion of the total 
inventory is large. 

Until such time as a value analysis is made and the valuable dollar ele- 
ments of the inventory are isolated, there is no firm basis for the con- 
trolling of inventories and the planning of production. The task is not easy 
and requires trained personnel with established objectives to accomplish 
the segregation and identification of items. Once this project has been ac- 
complished the benefits are immediate because policies can be established 
for handling the small number of valuable parts and the large number of 
low-value parts. 

This approach gets to the root of the excessive paperwork and immedi- 
ately eliminates a sizable volume of transactions. This saves on clerical 
costs, machine rentals and indirect shop labor. It directs the attention of 
all the organization to the valuable elements which should be most in- 
tensely studied and what action should be taken to improve control, or- 
dering, receipt, and to effect cost reductions through improvements in de- 
sign, methods, tooling and competitive prices, from vendors. It shows the 
need for condensing and standardizing of material, parts and products be- 
cause of the high number of items which are low value because of low 

This analysis has been made by some companies. It is known as the 
"A-B-C" or proportional parts analysis system. Its interesting implications 
can be seen from actual company experiences where it has been found 

Inventory Control and Production Planning 205 

that 10% of the inventory items account for approximately 75% of the 
dollar value. If this percentage is expanded to 25% of the items, it will 
usually cover 95% of the value, leaving 75% of the items accounting for 
only 5% of the total dollar inventory. Its principle is based on control 
by exception with clearly established policies and procedures which 
recognize the difference in objectives. 


Different inventory policies are required for the valuable items of the 
inventory and for the other items. It may be necessary to have daily, 
weekly or monthly review of the valuable items and to reduce their in- 
vestment through continuous scheduling of receipts. The fixed costs of 
order-writing, follow-up, set-up, material handling and record keeping 
are a small proportion of the total cost of the item because of the high 
usage and high value criterion. Changes in sales forecasts or trends will be 
quickly reflected in increases or decreases in the quantities ordered and 
the scheduling of receipts. The most qualified personnel should be selected 
to control these items. 

The effects on the factors of obsolescence and carrying costs are de- 
creased because they are based on a time element which is shorter in the 
case of these valuable parts. This is usually not considered in the various 
formulas for determining the ordering quantities for parts since they are 
based on a straight-line computation for obsolescence and carrying 

One of the difficulties encountered in management policies relating to 
close inventory control is the interpretation placed on such directives by 
the personnel responsible for their execution. When management desires 
an inventory reduction it wants dollars cut out of the inventory but does 
not want indirect labor, direct labor, or handling and set-up costs to be 
materially increased. The actual results of many inventory directives are 
that ordering quantities are reduced for both "A" and "B-C" items. The 
intent was to provide for increased control on the valuable items "A," in 
the inventory where the dollar inventory reductions can be achieved. 

By increasing the frequency of order issuance for all items a flood of 
orders is created by the purchasing and production planning personnel. 
This quickly results in high costs of set-up, paperwork, material handling, 
tool and die maintenance and handling, poor relationships with vendors 
and lack of parts to meet production schedules. It leads into excessive 
expediting, breaking into machines, and hurry-up calls to purchasing (and 
manufacturing) to get additional parts or material. The costs of store- 
room operations go up since the frequency of handling and the amount of 
storage space is increased. 

20 5 Inventory and the Administrator 

The keystone to inventory policy is determining of the proper ordering 
quantity. Some general guides can be laid down. These are illustrative 
and would vary with the industry, the type of product, and the labor and 
material contents of the piece. 

i. High- Value Items ("A" 10% of items)-These would be ordered a maxi- 
mum of six times a year and a minimum of three times, with a dollar limit of 
$300.00 per order. 

2. Middle- Value Items ("B" 15% of items) -These would be ordered a 
maximum of three times a year and a minimum of two times a year, with a 
dollar limit of $150.00. 

3. Low- Value Items ("C" 75% of items)-These would be ordered a maxi- 
mum of two times a year and a minimum of once a year, with a dollar limit 
of $80.00. 

From the above example, it can be seen that more months of supply can 
be purchased or manufactured within the dollar limits of the low-value 
item, and that the frequency of ordering is deliberately made greater in 
the high-value item. These various points can be covered in the respective 
inventory policies for the different classes of parts and will be easily under- 
stood by the people working with these instructions. Exceptions to the 
policy will occur because of a special situation for an item, but such devia- 
tions are approved by a responsible, major supervisor having the necessary 


Policies for lead times of high-value and low- value items can be used to 
gain cost and delivery advantages and to provide production flexibility for 
each class of stock parts and parts made to customer order which require 
the shortest possible cycle. 

Since the objective is for quick turnover and quick production of the 
valuable items, their lead time can be 15 to 30 days less than the lead time 
for the other stock items. The lead time for the special parts made to cus- 
tomer order could be 1 5 days less than the valuable items. This establishes 
a priority of work which can easily be explained to shop supervision and 
also to the purchasing personnel. 

This recognizes that the low-value items are the most costly to produce 
in terms of their fixed cost, paperwork, handling and set-ups and it is im- 
portant to achieve economies wherever possible. Through the use of 
longer lead times for this class of item, the factors can use these items for 
fill-in work, for work planning and for utilization of material. The longer 
lead time permits this planning and grouping for savings in set-up and 
material without interference with the valuable items and the customer 

Inventory Control and Production Planning 207 

The attention of the organization is given to the high-value items to 
see that they are manufactured and purchased without any delay in either 
the paperwork, the procurement or the receiving of these items, and that 
the inventory investment is minimized. 

In accomplishing this objective, the organization can study the high- 
value items for possible short cuts in production methods or purchasing 
techniques. Any savings in lead time on these items has a marked effect 
on costs and on inventory investment. 

Provision is made for the quickest possible purchasing or manufacturing 
of special nonstock parts ahead of all stock parts. 


The segregation of the items into the high-value class helps the fore- 
casting of requirements for manpower, machines and material. These ele- 
ments can be projected for the relatively few number of valuable items 
and then a factor added to cover the rest of the items. This gives a good 
yardstick for personnel, machines and material requirements. 

Many attempts to quickly and currently provide such information have 
been defeated by including all items without the identification of the 
valuable elements of the inventory. The sheer volume of work involved 
caused the effort to fail and rendered any figures meaningless. 


The determination of an adequate protective balance to be carried for 
each stock item is essential. This quantity is insurance against stock deple- 
tion due to late deliveries from suppliers or from the factory and for un- 
usual sales which exceed the projected average pattern on which the or- 
dering quantity is based. 

The careful provision for such stock minimizes rush orders having to be 
placed to build up the supply. It prevents shortages developing which 
interfere with shipments or with assembly production. The breaking into 
production or into purchasing for these rush orders causes considerable 
lost time, expediting and other expensive operations. The performance, 
efficiency and production of the assembly department are seriously ham- 
pered. Excessive costs and customer complaints are incurred in correspond- 
ing with customers relative to failure to deliver on schedule. Often the 
low-value items requiring a minimum of dollar investment are the ones 
which are out of stock and are delaying shipment and limiting inventory 

This protective stock can also be used to provide for uninterrupted 
service in the event of strikes, especially in the case of difficult-to-procure 

208 Inventory and the Administrator 

material and in vital raw materials which are basic to the production of 
the company. 

Cash carefully invested in this reserve will yield a major return. 

General rules can be established for the protective stock by class of 
part. A protective stock for valuable items might be one-half a month's 
projected average sales and for other items two months' projected average 
sales. Further refinements can be determined for the individual part by 
analyzing each part to show its average delivery times, average sales, and 
the deviation from the average sale. This deviation can be established as 
the protective stock. This will provide coverage for this quantity and can 
result in less inventory. 

The analysis can be done on the valuable items. On the others it would 
not be necessary to do such an analysis, both because of limited inventory 
investment and because of the wide fluctuations which usually occur in 
these parts. 


In many situations a careful examination will show that it is advan- 
tageous to carry inventory for a period of time in order to use the slow 
sales period to produce parts, subassemblies or final units. This will depend 
on the situation involved. 

It is possible to minimize this type of investment and the required stor- 
age space by keeping the part and subassemblies in stock and increase the 
production on the assembly lines by adding men through transfer from 
parts departments or by part-time help. This avoids tying up assembly 
labor in the inventory. 

Another approach is to assemble all manufactured parts but leave out 
the critical purchased items, such as an expensive motor or an expensive 
exhaust installation. There are many ways of approaching this problem 
which can be adopted by the individual company to meet its needs. 

The planned production of such inventory by trained workers avoids 
the costs of training new personnel, the attendant risk of poor quality and 
of tool and die breakage, and the unfavorable element of having parts sub- 
contracted thus perhaps inviting competition into your business. Further 
savings are realized through reducing overtime payments and higher pro- 
duction because of normal working hours. 

It is important that the working force be kept flexible and trained to 
perform various jobs in order that the changing load picture can be met 
without interruption and delay, or without adding to the work force. 

The inception of such a program designed to train a diversified labor 
pool is expensive and requires a good training program under the direction 
of capable company personnel. 

Inventory Control and Production Planning 209 

The program helps in covering vacation periods and leads to the dis- 
covery of capabilities in personnel previously not observed. It gives added 
interest and stimulation to company personnel since it relieves the mo- 
notony of one job and one operation. 

The effect on personnel relations is good because it provides steady 
work, avoids high costs and unsatisfactory production resulting from the 
bumping of employes from job to job to adjust to the new work force 
and the company seniority. This factor alone is an inducement to carefully 
planning for an even flow of production and an increased inventory in- 
vestment for a limited period of time. 

Having an adequate and planned inventory before the seasonal upsurge 
permits the company to take unusual orders in the busy season without 
undue difficulties and extension of work hours or of work force. 

Many companies are finding that it is sound business to borrow money 
for inventories built for this purpose. The costs and penalties of intermit- 
tent production represent a very high operating charge. Any decision 
regarding major inventory investment is a responsibility of top manage- 
ment. Such a policy should be fully developed, formalized, communicated 
to its administrators, and be carefully checked at regular intervals. This 
program should have dollar limits for the various inventory levels which 
will reflect the program authorized by management. 


Separate policies for the manufacturing of parts, the purchasing of parts 
and the assembling of purchased and manufactured parts into sub- or 
final assemblies will provide flexibility of operations, will limit inventory 
investment, and will enable the production load to be treated in its basic 

Each manufactured part has its own characteristics and will be ordered 
accordingly. A plan can be developed for carrying out this objective by 
having the "A" parts ordered in greater quantities during the seasonal lull, 
which automatically increases the reserve stock available to meet the sea- 
sonal peaks which are reflected in the increased assembly schedule. A re- 
view can also be made of "B" and "C" parts and orders placed for those 
items which are approaching their reorder point. This saves machine time 
in the busy season and helps delivery to customers. 

A separate policy is required for purchased parts since there is usually 
no advantage to having them received ahead of the time when they are re- 
quired for final assemblies. The purchasing department can carefully an- 
alyze and schedule the receipts of such parts with the vendors on such a 
basis as to have the vendor protected for the necessary inventory of raw 
materials which might be required for the final purchased parts. This keeps 

2 io Inventory and the Administrator 

the inventory investment at a minimum and yet affords protection to the 
vender in order to guarantee a smooth flow of purchased parts in the busy 
season. Many vendors are willing to carry a stock of finished parts built 
ahead in order that they may level their productions. The vendor retains 
title to the parts and is protected since he knows the anticipated usage 
schedules of the customer. 

The policy for the control of subassemblies and final assemblies should 
be carefully developed because of their effect on inventory investment, 
the flexibility of parts, and the storage problems created. Generally speak- 
ing, it is easier to increase the production tempo of assembly lines than the 
production of parts. Under modern production methods the number of 
stations on the line can be increased and additional subassembly lines can 
be created to feed into the main assembly line. 

A careful review should be made of the production drawings and pro- 
duction practices relating to stocked subassemblies. It is important that 
they be kept to a minimum. Properly controlled, they can be used for 
building up stock ahead of the assembly line if the consideration of storage 
and value meets the tests of such practice. In many situations it is better 
to treat them as sublines which are started earlier than the main assembly 
lines and are feeder lines. In this situation a bank of such line assemblies 
is created ahead of the main line. 

Good policies and practices relative to stocked subassemblies, line sub- 
assemblies and final assemblies are the result of close and careful considera- 
tions on the part of the engineering, industrial engineering, production 
planning and production supervisory personnel, all having a clearly de- 
fined objective upon which to base their composite decision. All their 
knowledge is required to develop and reduce to practice sound policies for 
the control of assemblies. 


A smooth flow of production is difficult to obtain without good rela- 
tionships with vendors. Policies should be developed which will interest 
good vendors in establishing— and maintaining with the customer— records 
of reliability of delivery, good quality of merchandise, and a fair price 
structure. In return, they should expect reasonable requests for deliveries 
with consideration given to good ordering quantities. Suppliers do not 
expect to have purchasing orders placed for a year's supply of the valuable 
items. On the other hand, they do expect and encourage ordering of low- 
value parts on as infrequent a basis as possible. 

They will appreciate being furnished projection of anticipated require- 
ments with progressive segments of the projection being firmed up as 
commitments. This enables them to make their own company plans on a 
better basis. In many situations they are willing to carry some stock if it 

Inventory Control and Production Planning 2 1 1 

can be proven that they will have steady and profitable relationships with 
the customer. 

Any of the above advantages can be achieved by carefully developed 
production programs and by interchange of information between the 
vendor and the customer. In many situations the vendor can make sug- 
gestions as to improvement in design specifications, which will permit 
price reduction by the vendor without sacrifice in product acceptability. 
This calls for close relationship between the engineering, manufacturing, 
and purchasing personnel and the vendor. 

The records of the purchasing department should be established to 
quickly reveal unreliable vendors whose quality, price or deliveries are un- 
satisfactory. Such records should be used to correct the situation by satis- 
factory performance or by the changing of vendors. 

A company policy can be developed which provides for at least two 
vendors as sources. This provides protection against interruption in the 
supply and also develops competition between the vendors. Such a policy 
should be explained to the vendors and requires the engineering specifica- 
tions being written to permit such practice. 


Programs designed for a steady flow of production require good store- 
rooms and good storeroom procedures. Policies should be developed for 
having well-protected and controlled storage areas, for the storing of parts 
having similar characteristics in similar storage equipment, for the receipt 
and issue of parts using the proper transaction paper, and for daily opera- 
tions being kept current for receipts and issues. 

Characteristics of parts as to storage and issue should be studied. Pro- 
vision should be made for simplifying the handling of receipts and issues, 
giving consideration to simplification of paper transactions on inventory 
and production records. This can result in substantial improvements and 
savings in clerical and storeroom costs. The time required for the taking 
of inventories can be considerably shortened through techniques of stor- 
age, stores control, and inventory and production records. 

Examples of practices which result in over-all savings are the issuing of 
working supplies of hardware to assembly lines, the bundling, tagging and 
issuing of parts in bulk lots, the designing of racks and handling equipment 
for special classes of items, and the purchasing of packaged lots from ven- 
dors which fit the production requirements of the company. 

Working hours in the storerooms can be established to provide a steady 
flow of parts to assembly floors while, at the same time, making sure that 
receipts of parts from the manufacturing and purchasing departments are 
being promptly counted, reported and placed in their permanent locations. 

2i2 Inventory and the Administrator 

Interruptions in regular work can be minimized by having regular hours 
of issues with the cooperation of the production supervisors. 

Provision for adequate storage and personnel is very important to the 
whole program of production flow. As much consideration should be 
given to the adequacy of storage space, equipment and personnel as is 
given to manufacturing space and personnel. Inadequate storage space 
will result in manufacturing space being used for storage, and also in un- 
satisfactory control of stores property with the attendant delays in pro- 


Policies relating to the changing of models or parts and their timing 
have a major effect on deliveries, costs and manpower, machine and ma- 
terial utilization. It is reasonable to assume that this major phase will re- 
ceive increased attention in the future. So many different phases of the 
business are affected that improperly timed changes cause strains on the 
entire organization. 

Policies in this field should provide for long-range thinking on product 
development, for the detailed scheduling of such projects so that they are 
placed in production in off-season times, that changes to parts be carefully 
studied in relation to value of the part ("A" parts should receive major 
attention), and that careful consideration be given to existing tooling, 
stock on hand, and effect on service as to interchangeability. 

Any decisions involving changes to parts or products should be 
promptly conveyed to the personnel responsible for production and in- 
ventory. This permits them to take immediate steps to prevent the ac- 
cumulation of stock for parts which are to be obsoleted and allows them 
time to obtain new material, tooling and other items required for the 
proposed change. 

As the development proceeds, provisions should be made for seeing that 
inventory policies are carried out in the creation of assemblies, subassem- 
blies and parts which follow inventory and production practices. The 
production drawings for assemblies and parts should reflect such policies 
since they are the source required to make these policies effective. This 
requires a thorough understanding on the part of the engineering, in- 
dustrial engineering, production planning and production supervision as 
to the over-all policies of the company related to inventory and produc- 
tion philosophy. The reduction of the product development to the pro- 
duction design affords the opportunity to carry out such policies. 

Proper scheduling of work will permit the engineering personnel being 
given an opportunity to level their work load through working on de- 
velopment, design and engineering change in periods not involved on 
heavy work for customer application. 

Inventory Control and Production Planning 2 1 3 


Careful and accurate planning of inventories and production is based 
on reliable sales forecasts. The expected trends of business are guideposts 
to point the way to decisions for short- and long-range programs. 

The minimum amount of forecasting which is necessary is an annual 
forecast which is firmed up on a quarterly basis. The quarterly figures are 
used for commitments and manufacturing schedules. The annual forecast 
affects the annual ordering quantities, the expected commitments with 
vendors, and the manpower and machines required to meet the anticipated 

It is usually helpful if the quarterly trend is a continuous three months' 
projection because action can be taken to provide for the trend indicated 
in the newly added month. 

The annual forecasts should be prepared early enough to avoid having a 
gap in the long-term trends when the end of the current annual forecast 
is approached. 

These sales forecasts cannot be detailed to show the expected usage of 
the individual parts or subassemblies. These are controlled on the basis 
of the over-all indicated trend which, in turn, is reflected in the projected 
average monthly usage. This is used to determine the ordering and control 
of the individual part. Usually the lead time of the part is sufficiently short 
as to enable changes in usage trends being reflected in the ordering of the 
part when it reaches its reviewing and ordering point. 

The valuable items ("A") should be closely checked for changes in pro- 
jected average monthly usage and full attention be given to those parts. 
Outstanding purchase or production orders should be adjusted to meet 
indicated changes in usage. This adjusts supply to the expected demand 
and keeps the dollar inventory in the optimum balance. Conversely, the 
less valuable items ("B" and "C") have little effect on the inventory dol- 
lars and changes in trends can be covered when the reviewing and order- 
ing point is reached. Under this concept the degree of attention and 
changes in order quantities and delivery date corresponds with the degree 
of value and investment. 


One of the greatest assets in any business is its relationships with its 
customers. A good part of this asset depends on the ability to make ship- 
ment on schedule, at a fair price and with good quality merchandise. 

The customers require quicker delivery when the supply situation is 
adequate because they will usually prefer having the company act as their 
warehouse and avoid cash outlay for inventory not currently being used. 
These factors are increasingly apparent in the present situation and the 

2 1 a Inventory and the Administrator 

only difference is the degree between various industries. Even though a 
specific industry at the moment is in a tight supply situation, there is no 
guarantee that that will continue forever. Accordingly, it is important that 
management prepare for that day by maintaining balanced, strategically 
located stocks of inventory available for quick shipment. The company 
which achieves these objectives has a distinct competitive advantage over 
others not so forward thinking. 

A concerted program will be required on the part of general, sales, fi- 
nancial, purchasing and manufacturing management since all will partici- 
pate in achieving the vital objective of customer goodwill and the addition 
of new customers who are attracted by progressive inventory management 
and reliability of source along with factors of price and quality. Special 
attention should be given to the allocation of stock to incoming orders 
from customers to avoid temporarily being out of stock because of filling 
a very large order when a partial shipment might have been made which 
would leave enough stock for the other customers. The decision and con- 
trol of finished stock should generally be assigned to the merchandising 
group, since they are most familiar with their customers and their require- 
ments. Some organizations have set up inventory controllers within the 
merchandising group for this purpose, as well as tying in with general 
management policies on inventory levels. 

One of the major tests of a good inventory control and production plan- 
ning system is its ability to meet customer requirements based on good 
sales forecasting and good procurement and manufacturing practices. 
Usually a company having a high degree of reliability in meeting shipping 
promises will also have high profits within the industry. 


John F. Magee* 

Mr. John F. Magee examines the functions of an inventory and discusses 
the conflicting attitudes influencing the, size and structure of the inventory. 
Among the problems discussed is the difficulty of finding the proper costs 
to assign to the maintenance of an inventory as well as other matters to 
be taken into account in establishing a proper policy. An illustration of a 
hypothetical inventory situation is used to tie together the main points 
covered in the first of a three article series. 

* From Harvard Business Review, XXXIV, 1 (1956), 49-60. Reprinted by permis- 
sion of the Harvard Business Review. 

Guides to Inventory Policy I 215 

"Why are we always out of stock?" So goes the complaint of great 
numbers of businessmen faced with the dilemmas and frustrations of at- 
tempting simultaneously to maintain stable production operations, provide 
customers with adequate service, and keep investment in stocks and equip- 
ment at reasonable levels. 

But this is only one of the characteristic problems business managers 
face in dealing with production planning, scheduling, keeping inventories 
in hand, and expediting. Other questions-just as perplexing and baffling 
when managers approach them on the basis of intuition and pencil work 
alone-are: How often should we reorder, or how should we adjust pro- 
duction, when sales are uncertain? What capacity levels should we set for 
job-shop operations? How do we plan production and procurement for 
seasonal sales? And so on, and so on. 

In this series of articles, I will describe some of the technical develop- 
ments which aim at giving the business manager better control over in- 
ventory and scheduling policy. While these techniques sometimes em- 
ploy concepts and language foreign to the line executive, they are far 
from being either academic exercises or mere clerical devices. They are de- 
signed to help the business manager make better policy decisions and get 
his people to follow policy more closely. 

In the present article major attention will be devoted to (a) the con- 
ceptual framework of the analytic approach, including the definition of 
inventory function and the measurement of operational costs; and (b) 
the problem of optimum lot size, with a detailed case illustration showing 
how the techniques are applied. 

This case reveals that the appropriate order quantity and the average 
inventory maintained do not vary directly with sales, and that a good 
answer to the lot size question can be obtained with fairly crude cost 
data, provided that a sound analytical approach is used. The case also 
shows that the businessman does not need calculus to solve many inven- 
tory problems (although use has to be made of it when certain complica- 
tions arise) . 


The question before management is: How big should inventories be? 
The answer to this is obvious-they should be just big enough. But what 
is big enough? 

This question is made more difficult by the fact that generally each in- 
dividual within a management group tends to answer the question from his 
own point of" view. He fails to recognize costs outside his usual frame- 
work. He tends to think of inventories in isolation from other operations. 
The sales manager commonly says that the company must never make a 

2 1 6 Inventory and the Administrator 

customer wait; the production manager says there must be long manufac- 
turing runs for lower costs and steady employment; the treasurer says that 
large inventories are draining off cash which could be used to make a 

Such a situation occurs all the time. The task of all production planning, 
scheduling, or control functions, in fact, is typically to balance conflicting 
objectives such as those of minimum purchase or production cost, mini- 
mum inventory investment, minimum storage and distribution cost, and 
maximum service to customers. 


Fundamentally, inventories serve to uncouple successive operations in 
the process of making a product and getting it to consumers. For exam- 
ple, inventories make it possible to process a product at a distance from 
customers or from raw material supplies, or to do two operations at a 
distance from one another (perhaps only across the plant). Inventories 
make it unnecessary to gear production directly to consumption or, al- 
ternatively, to force consumption to adapt to the necessities of production. 
In these and similar ways, inventories free one stage in the production- 
distribution process from the next, permitting each to operate more eco- 

The essential question is: At what point does the uncoupling function 
of inventory stop earning enough advantage to justify the investment re- 
quired? To arrive at a satisfactory answer we must first distinguish be- 
tween (a) inventories necessary because it takes time to complete an 
operation and to move the product from one stage to another; and (b) 
inventories employed for organizational reasons, i.e., to let one unit sched- 
ule its operations more or less independently of another. 


Inventory balances needed because of the time required to move stocks 
from one place to another are often not recognized, or are confused with 
inventories resulting from other needs— e.g., economical shipping quantities 
(to be discussed in a later section). 

The average amount of movement inventory can be determined from 
the mathematical expression / = 5 X T in which S represents the average 
sales rate, T the transit time from one stage to the next, and / the move- 
ment inventory needed. For example, if it takes two weeks to move ma- 
terials from the plant to a warehouse, and the warehouse sells ioo units per 
week, the average inventory in movement is ioo units per week times i 
weeks, or 200 units. From a different point of view, when a unit is manu- 
factured and ready for use at the plant, it must sit idle for two weeks while 

Guides to Inventory Policy I 217 

being moved to the next station (the warehouse); so, on the average, 
stocks equal to two weeks' sales will be in movement. 

Movement inventories are usually thought of in connection with move- 
ment between distant points— plant to warehouse. However, any plant 
may contain substantial stocks in movement from one operation to an- 
other—for example, the product moving along an assembly line. Move- 
ment stock is one component of the "float" or in-process inventory in a 
manufacturing operation. 

The amount of movement stock changes only when sales or the time in 
transit is changed. Time in transit is largely a result of method of trans- 
portation, although improvements in loading or dispatching practices may 
cut transit time by eliminating unnecessary delays. Other somewhat more 
subtle influences of time in transit on total inventories will be described in 
connection with safety stocks. 


Management's most difficult problems are with the inventories that 
"buy" organization in the sense that the more of them management carries 
between stages in the manufacturing-distribution process, the less coordi- 
nation is required to keep the process running smoothly. Contrariwise, if 
inventories are already being used efficiently, they can be cut only at the 
expense of greater organization effort— e.g., greater scheduling effort to 
keep successive stages in balance, and greater expediting effort to work 
out of the difficulties which unforeseen disruptions at one point or another 
may cause in the whole process. 

Despite superficial differences among businesses in the nature and char- 
acteristics of the organization inventory they maintain, the following 
three functions are basic: 

1. Lot size inventories are probably the most common in business. They 
are maintained wherever the user makes or purchases material in larger lots 
than are needed for his immediate purposes. For example, it is common prac- 
tice to buy raw materials in relatively large quantities to order to obtain quan- 
tity price discounts, keep shipping costs in balance, and hold down clerical 
costs connected with making out requisitions, checking receipts, and handling 
accounts payable. Similar reasons lead to long production runs on equipment 
calling for expensive setup, or to sizable replenishment orders placed on fac- 
tories by field warehouses. 

2. Fluctuation stocks, also very common in business, are held to cushion 
the shocks arising basically from unpredictable fluctuations in consumer de- 
mand. For example, warehouses and retail outlets maintain stocks to be able to 
supply consumers on demand, even when the rate of consumer demand may 
show quite irregular and unpredictable fluctuations. In turn, factories maintain 
stocks to be in a position to replenish retail and field warehouse stocks in line 
with customer demands. 

2 1 8 Inventory and the Administrator 

Short-term fluctuations in the mix of orders on a plant often make it neces- 
sary to carry stocks of parts of subassemblies, in order to give assembly opera- 
tions flexibility in meeting orders as they arise while freeing earlier operations 
(e.g., machining) from the need to make momentary adjustments in schedules 
to meet assembly requirements. Fluctuation stocks may also be carried in 
semifinished form in order to balance out the load among manufacturing de- 
partments when orders received during the current day, week, or month may 
put a load on individual departments which is out of balance with long-run re- 

In most cases, anticipating all fluctuations is uneconomical, if not impossible. 
But a business cannot get along without some fluctuation stocks unless it is 
willing and able always to make its customers wait until the material needed 
can be purchased conveniently or until their orders can be scheduled into pro- 
duction conveniently. Fluctuation stocks are part of the price we pay for our 
general business philosophy of serving the consumers' wants (and whims!) 
rather than having them take what they can get. 

3. Anticipation stocks are needed where goods or materials are consumed 
on a predictable but changing pattern through the year, and where it is desir- 
able to absorb some of these changes by building and depleting inventories 
rather than by changing production rates with attendant fluctuations in em- 
ployment and additional capital capacity requirements. For example, inven- 
tories may be built up in anticipation of a special sale or to fill needs during a 
plant shutdown. 

The need for seasonal stocks may also arise where materials (e.g., agricul- 
tural products) are produced at seasonally fluctuating rates but where con- 
sumption is reasonably uniform; here the problems connected with producing 
and storing tomato catsup are a prime example. 


The joker is that the gains which these organization inventories achieve 
in the way of less need for coordination and planning, less clerical effort 
to handle orders, and greater economies in manufacturing and shipping are 
not in direct proportion to the size of inventory. Even if the additional 
stocks are kept well balanced and properly located, the gains become 
smaller, while at the same time the warehouse, obsolescence, and capital 
costs associated with maintaining inventories rise in proportion to, or 
perhaps even at a faster rate than, the inventories themselves. To illustrate: 

Suppose a plant needs 2,000 units of a specially machined part in a year. If 
these are made in runs of 100 units each, then 20 runs with attendant setup 
costs will be required each year. 

If the production quantity were increased from 100 to 200 units, only 10 
runs would be required— a 50% reduction in setup costs, but a 100% increase 
in the size of a run and in the resulting inventory balance carried. 

If the runs were further increased in length to 400 units each, only 5 produc- 
tion runs during the year would be required— only 25% more reduction in 
setup costs, but 200% more increase in run length and inventory balances. 

The basic problem of inventory policy connected with the three types 
of inventories which "buy" organization is to strike a balance between 

Guides to Inventory Policy I 219 

the increasing costs and the declining return earned from additional stocks. 
It is because striking this balance is easier to say than to do, and because it 
is a problem that defies solution through an intuitive understanding alone, 
that the new analytical concepts are necessary. 


This brings us face to face with the question of the costs that influence 
inventory policy, and the fact, noted earlier, that they are characteris- 
tically not those recorded, at least not in directly available form, in the 
usual industrial accounting system. Accounting costs are derived under 
principles developed over many years and strongly influenced by tradition. 
The specific methods and degree of skill and refinement may be better in 
particular companies, but in all of them the basic objective of accounting 
procedures is to provide a fair, consistent, and conservative valuation of 
assets and a picture of the flow of values in the business. 

In contrast to the principles and search for consistency underlying ac- 
counting costs, the definition of costs for production and inventory con- 
trol will vary from time to time-even in the same company-according to 
the circumstances and the length of the period being planned for. The 
following criteria apply: 

1. The costs shall represent "out-of-pocket" expenditures, i.e., cash actually 
paid out or opportunities for profit foregone. Overtime premium payments are 
out-of-pocket; depreciation on equipment on hand is not. To the extent that 
storage space is available and cannot be used for other productive purposes, no 
out-of-pocket cost of space is incurred; but to the extent that storage space is 
rented (out-of-pocket) or could be used for other productive purposes (fore- 
gone opportunity), a suitable charge is justified. The charge for investment is 
based on the out-of-pocket investment in inventories or added facilities, not on 
the "book" or accounting value of the investment. 

The rate of interest charged on out-of-pocket investment may be based 
either on the rate paid banks (out-of-pocket) or on the rate of profit that 
might reasonably be earned by alternative uses of investment (foregone oppor- 
tunity), depending on the financial policies of the business. In some cases, a 
bank rate may be used on short-term seasonal inventories and an internal rate 
for long-term, minimum requirements. 

Obviously, much depends on the time scale in classifying a given item. In 
the short run, few costs are controllable out-of-pocket costs; in the long run, 
all are. 

2. The costs shall represent only those out-of-pocket expenditures or fore- 
gone opportunities for profit whose magnitude is affected by the schedule or 
plan. Many overhead costs, such as supervision costs, are out-of-pocket, but 
neither the timing nor the size is affected by the schedule. Normal material 
and direct labor costs are unaffected in total and so are not considered directly; 
however, these as well as some components of overhead cost do represent out- 
of-pocket investments, and accordingly enter the picture indirectly through 
any charge for capital. 

2 2 o Inventory and the Administrator 


Among the costs which directly influence inventory policy are (a) 
costs depending on the amount ordered, (b) production costs, and (c) 
costs of storing and handling inventory. 

Costs that depend on the amount ordered— These include, for example, 
quantity discounts offered by vendors; setup costs in internal manufacturing 
operations and clerical costs of making out a purchase order; and, when ca- 
pacity is pressed, the profit on production lost during downtime for setup. 
Shipping costs represent another factor to the extent that they influence the 
quantity of raw materials purchased and resulting raw stock levels, the size of 
intraplant or plant-warehouse shipments, or the size and the frequency of 
shipments to customers. 

Production costs— Beyond setup or change-over costs, which are included in 
the preceding category, there are the abnormal or nonroutine costs of produc- 
tion whose size may be affected by the policies or control methods used. 
(Normal or standard raw material and direct labor costs are not significant in 
inventory control; these relate to the total quantity sold rather than to the 
amount stocked.) Overtime, shakedown, hiring, and training represent costs 
that have a direct bearing on inventory policy. 

To illustrate, shakedown or learning costs show up wherever output during 
the early part of a new run is below standard in quantity or quality. A cost of 
undercapacity operation may also be encountered— for example, where a basic 
labor force must be maintained regardless of volume (although sometimes this 
can be looked on as part of the fixed facility cost, despite the fact that it is 
accounted for as a directly variable labor cost). 

Costs of handling and storing inventory— In this group of costs affected by 
control methods and inventory policies are expenses of handling products in 
and out of stock, storage costs such as rent and heat, insurance and taxes, 
obsolescence and spoilage costs, and capital costs (which will receive detailed 
examination in the next section). 

Inventory obsolescence and spoilage costs may take several forms, including 
(i) outright spoilage after a more or less fixed period; (2) risk that a particu- 
lar unit in stock or a particular product number will (a) become technolog- 
ically unsalable, except perhaps at a discount or as spare parts, (b) go out of 
style, or (c) spoil. 


Evaluating the effect of inventory and scheduling policy upon capital 
investment and the worth of capital tied up in inventories is one of the 
most difficult problems in resolving inventory policy questions. 

Think for a moment of the amount of capital invested in inventory. 
This is the out-of-pocket, or avoidable, cash cost for material, labor, and 
overhead of goods in inventory (as distinguished from the "book" or ac- 
counting value of inventory). For example, raw materials are normally 
purchased in accordance with production schedules; and if the production 
of an item can be postponed, buying and paying for raw materials can 
likewise be put off. 

Guides to Inventory Policy I 221 

Usually, then, the raw material cost component represents a part of the 
out-of-pocket inventory investment in finished goods. However, if raw 
materials must be purchased when available (e.g., agricultural crops) re- 
gardless of the production schedule, the raw material component of 
finished product cost does not represent avoidable investment and there- 
fore should be struck from the computation of inventory value for plan- 
ning purposes. 

As for maintenance and similar factory overhead items, they are usually 
paid for the year round, regardless of the timing of production scheduled; 
therefore these elements of burden should not be counted as part of the 
product investment for planning purposes. (One exception: if, as some- 
times happens, the maintenance costs actually vary directly with the pro- 
duction rate as, for example, in the case of supplies, they should of course 
be included.) 

Again, supervision, at least general supervision, is usually a fixed monthly 
cost which the schedule will not influence, and hence should not be in- 
cluded. Depreciation is another type of burden item representing a charge 
for equipment and facilities already bought and paid for; the timing of 
the production schedule cannot influence these past investments and, 
while they represent a legitimate cost for accounting purposes, they should 
not be counted as part of the inventory investment for inventory and 
production planning purposes. 

In sum, the rule is this: for production planning and inventory manage- 
ment purposes, the investment value of goods in inventory should be 
taken as the cash outlay made at the time of production that could have 
been delayed if the goods were not made then but at a later time, closer to 
the time of sale. 

Cost of Capital Invested. This item is the product of three factors: (a) 
the capital value of a unit of inventory, (b) the time a unit of product is in 
inventory, and (c) the charge or imputed interest rate placed against a 
dollar of invested cash. The first factor was mentioned above. As for the 
second, it is fixed by management's inventory policy decisions. But these 
decisions can be made economically only in view of the third factor. 
This factor depends directly on the financial policy of the business. 

Sometimes businessmen make the mistake of thinking that cash tied up 
in inventories costs nothing, especially if the cash to finance inventory is 
generated internally through profits and depreciation. However, this im- 
plies that the cash in inventories would otherwise sit idle. In fact, the cash 
could, at least, be invested in government bonds if not in inventories. And 
if it were really idle, the cash very likely should be released to stock- 
holders for profitable investment elsewhere. 

Moreover, it is dangerous to assume that, as a "short-term" investment, 
inventory is relatively liquid and riskless. Businessmen say, "After all, we 

222 Inventory and the Administrator 

turn our inventory investment over six times a year." But, in reality, in- 
ventory investment may or may not be short-term and riskless, depending 
on circumstances. No broad generalization is possible, and each case must 
be decided on its own merits. 

Finally, it might be pointed out that the cost of the dollars invested in 
inventory may be underestimated if bank interest rate is used as the basis, 
ignoring the risk-bearing or entrepreneur's compensation. How many 
businessmen are actually satisfied with uses of their companies' capital 
funds which do not earn more than a lender's rate of return? In choosing 
a truly appropriate rate— a matter of financial policy— the executive must 
answer some questions: 

i. Where is the cash coming from— inside earnings or outside financing? 

2. What else could we do with the funds, and what could we earn? 

3. When can we get the investment back out, if ever? 

4. How much risk of sales disappointment and obsolescence is really con- 
nected with this inventory? 

5. How much of a return do we want, in view of what we could earn else- 
where or in view of the cost of money to us and the risk the inventory invest- 
ment entails? 

Investment in Facilities. Valuation of investment in facilities is generally 
important only in long-run planning problems— as, for example, when in- 
creases in productive or warehouse capacity are being considered. (Where 
facilities already exist and are not usable for other purposes, and where 
planning or scheduling do not contemplate changing these existing facili- 
ties, investment is not affected.) 

Facilities investment may also be important where productive capacity 
is taxed, and where the form of the plan or schedule will determine the 
amount of added capacity which must be installed, either to meet the 
plan itself or for alternative uses. In such cases, considerable care is 
necessary in defining the facilities investment in order to be consistent 
with the principles noted above: i.e., that facilities investment should repre- 
sent out-of-pocket investment, or, alternatively, foregone opportunities to 
make out-of-pocket investment elsewhere. 


An important objective in most production planning and inventory 
control systems is maintenance of reasonable customer service. An evalu- 
ation of the worth of customer service, or the loss suffered through poor 
service, is an important part of the problem of arriving at a reasonable in- 
ventory policy. This cost is typically very difficult to arrive at, including 
as it does the paper work costs of rehandling back orders and, usually 
much more important, the effect that dissatisfaction of customers may 
have on future profits. 

Guides to Inventory Policy 1 223 

In some cases it may be possible to limit consideration to the cost of 
producing the needed material on overtime or of purchasing it from the 
outside and losing the contribution to profit which it would have made. 
On the other hand, sometimes the possible loss of customers and their sales 
over a substantial time may outweigh the cost of direct loss in immediate 
business, and it may be necessary to arrive at a statement of a "reasonable" 
level of customer service— i.e., the degree of risk of running out of stock, 
or perhaps the number of times a year the management is willing to run 
out of an item. In other cases, it may be possible to arrive at a reasonable 
maximum level of sales which the company is prepared to meet with 
100% reliability, being reconciled to have service suffer if sales exceed 
this level. 

One of the uses of the analytic techniques described below and in fol- 
lowing parts of this series is to help management arrive at a realistic view 
of the cost of poor service, or of the value of building high service, by 
laying out clearly what the cost in inventory investment and schedule 
changes is to achieve this degree of customer service. Sometimes when 
these costs are clearly brought home, even a 100% service-minded man- 
agement is willing to settle for a more realistic, "excellent" service at mod- 
erate cost, instead of striving for "perfect" service entailing extreme cost. 


Now, with this background, let us examine in some detail one of the 
inventory problems which plague businessmen the most— that of the 
optimum size of lot to purchase or produce for stock. This happens also 
to be one of the oldest problems discussed in the industrial engineering 
texts-but this does not lessen the fact that it is one of the most profitable 
for a great many companies to attack today with new analytic techniques. 


This problem arises, as mentioned earlier, because of the need to pur- 
chase or produce in quantities greater than will be used or sold. Thus, 
specifically, businessmen buy raw materials in sizable quantities-carloads, 
or even trainloads-in order to reduce the costs connected with purchasing 
and control, to obtain a favorable price, and to minimize handling and 
transportation costs. They replenish factory in-process stocks of parts in 
sizable quantities to avoid, where possible, the costs of equipment setups 
and clerical routines. Likewise, finished stocks maintained in warehouses 
usually come in shipments substantially greater than the typical amount 
sold at once, the motive again being, in part, to avoid equipment setup 
and paper-work costs and, in the case of field warehouses, to minimize 
shipping costs. 

224 Inventory and the Administrator 

Where the same equipment is used for a variety of items, the equip- 
ment will be devoted first to one item and then to another in sequence, 
with the length of the run in any individual item to be chosen, as far as is 
economically possible, to minimize change-over cost from one item to 
another and to reduce the production time lost because of clean-out re- 
quirements during change-overs. Blocked operations of this sort are seen 
frequently, for example, in the petroleum industry, on packaging lines, 
or on assembly lines where change-over from one model to another may 
require adjustment in feed speeds and settings and change of components. 

In all these cases, the practicex)f replenishing stocks in sizable quantities 
compared with the typical usage quantity means that inventory has to be 
carried; it makes it possible to spread fixed costs (e.g., setup and clerical 
costs) over many units and thus to reduce the unit cost. However, one can 
carry this principle only so far, for if the replenishment orders become 
too large, the resulting inventories get out of line, and the capital and 
handling costs of carrying these inventories more than offset the possible 
savings in production, transportation, and clerical costs. Here is the matter, 
again, of striking a balance between these conflicting considerations. 

Even though formulas for selecting the optimum lot size are presented 
in many industrial engineering texts, few companies make any attempt to 
arrive at an explicit quantitative balance of inventory and change-over or 
setup costs. Why? 

For one thing, the cost elements which enter into an explicit solution 
frequently are very difficult to measure, or are only very hazily defined. 
For example, it may be possible to get a fairly accurate measure of the 
cost of setting up a particular machine, but it may be almost impossible 
to derive a precise measure of the cost of making out a new production 
order. Again, warehouse costs may be accumulated separately on the ac- 
counting records, but these rarely show that the cost of housing an ad- 
ditional unit of material may be. In my experience the capital cost, or 
imputed interest cost, connected with inventory investment never appears 
on the company's accounting records. 

Furthermore, the inventory is traditionally valued in such a way that 
the true incremental investment is difficult to measure for scheduling pur- 
poses. Oftentimes companies therefore attempt to strike only a qualitative 
balance of these costs to arrive at something like an optimum- or mini- 
mum-cost reorder quantity. 

Despite the difficulty in measuring costs— and indeed because of such 
difficulty— it is eminently worthwhile to look at the lot size problem ex- 
plicitly formulated. The value of an analytic solution does not rest solely 
on one's ability to plug in precise cost data to get an answer. An analytic 
solution often helps clarify questions of principle, even with only crude 
data available for use. 

Guides to Inventory Policy I 



To illustrate how the lot size problem can be attacked analytically— and 
what some of the problems and advantages of such an attack are— let us 
take a fictitious example. The situation is greatly oversimplified on purpose 
to get quickly to the heart of the analytic approach. 

Elements of the Problem. Brown and Brown, Inc., an automotive parts 
supplier, produces a simple patented electric switch on long-term con- 
tracts. The covering is purchased on the outside at $0.01 each, and 1,000 
are used regularly each day, 2 50 days per year. 

The casings are made in a nearby plant, and B. and B. sends its own 
truck to pick them up. The cost of truck operation, maintenance, and the 
driver amounts to $10 per trip. 

The company can send the truck once a day to bring back 1 ,000 casings 
for that day's requirements, but this makes the cost of a casing rather high. 
The truck can go less frequently, but this means that it has to bring back 
more than the company needs for its immediate day-to-day purposes. 

The characteristic "saw-tooth" inventory pattern which will result is 
shown in exhibit i, where 1,000 Q casings are picked up each trip (Q be- 


(1,000 casings obtained per replenishment trip; 1,000 casings used per day) 



1,000 Q 


ing whatever number of days' supply is obtained per replenishment trip). 
These are used up over a period of Q days. When the inventory is de- 
pleted again, another trip is made to pick up Q days' supply or 1,000 Q 
casings once more, and so on. 

2 2 ^ Inventory and the Administrator 

B. and B. estimates that the cost of storing casings under properly con- 
trolled humidity conditions is $i per 1,000 casings per year. The company 
wants to obtain a 10% return on its inventory investment of $10 (1,000 
times $0.01), which means that it should properly charge an additional $1 
(10% of $10), making a total inventory cost of $2 per 1,000 casings 
per year. 



300 T 

250 - 








» I ' 1 * 

20,000 40,000 60,000 80,000 100,000 

Quantity of casings obtained each trip 

(Note that, in order to avoid undue complications, the inventory in- 
vestment charge is made here only against the purchase price of the 
casings and not against the total delivery cost including transportation. 
Where transportation is a major component of total cost, it is of course 
possible and desirable to include it in the base for the inventory charge.) 

Guides to Inventory Policy I 227 

Graphic Solution. Brown and Brown, Inc., can find what it should do 
by means of a graph (see exhibit ii) showing the annual cost of buying, 
moving, and storing casings: 

The broken line shows total trucking costs versus the size of the individual 
purchase quantity: 

If 1,000 casings are purchased at a time, the total cost is $10 times 250 trips, 

or $2,500 per year. 

If 10,000 casings are purchased at one time, only 25 trips need be made, for 

a total cost of $250 per year. 

If 100,000 casings are purchased, only i x h trips, on the average, have to be 

taken each year, for a total cost of $25. 

The dotted line shows the inventory cost compared with the size of the 
purchased quantity: 

If 10,000 casings are purchased at one time, the inventory at purchase will 
contain 10,000, and it will gradually be depleted until none are on hand, 
when a new purchase will be made. The average inventory on hand thus will 
be 5,000 casings. The cost per year will be $2 times 5,000 casings, or $10. 



The total annual cost of supplying casings is equal to the sum of the 
direct cost of the casings, plus the trucking cost, plus the inventory and 
storage cost. 


T = total annual cost 

b = unit purchase price, $10 per 1,000 casings 

s — annual usage, 250,000 casings 

A = trucking cost, $10 per trip 

N — number of trips per year 

i = cost of carrying casings in inventory at the annual rate of $2 per 

1,000, or $0,002 per casing 
x — size of an individual purchase (x/i = average inventory) 

The basic equation will be: 

T — bs + AN + ix/i 
The problem is to choose the minimum-cost value of x (or, if desired, 
N). Since x is the same as s/N, N can be expressed as s'/x. Substituting 
s/x for N in the above equation, we get: 

T = bs + As/x + ix/i 
From this point on we shall use differential calculus. The derivative of total 
cost, T, with respect to x will be expressed as: 

dT/dx = -As/x 2 + i/i 


Inventory and the Administrator 

And the minimum-cost value of x is that for which the derivative of total 
cost with respect to x equals zero. This is true when: 

x — \/iAs/i 
Substituting the known values for A, s, and i: 

x — a/2 • io • 2 50,000/. 002 = 50,000 casings 

Similarly, if 100,000 casings are purchased at one time, the average inventory- 
will be 50,000 casings, and the total inventory and storage cost will be $100. 

The solid line is the total cost, including both trucking and inventory 
and storage costs. The total cost is at a minimum when 50,000 casings are 
purchased on each trip and 5 trips are made each year, for at this point 
the total trucking cost and the total inventory and storage cost are equal. 

The solution to B. and B.'s problem can be reached algebraically as well 
as graphically, exhibit hi shows how the approach works in this very 
simple case. 


The problem of Brown and Brown, Inc., though artificial, is not too far 
from the questions many businesses face in fixing reorder quantities. 

Despite the simplifications introduced— for example, the assumption 
that usage is known in advance— the method of solution has been found 
widely useful in industries ranging from mail order merchandising (re- 
plenishing staple lines), through electrical equipment manufacturing (or- 
dering machined parts to replenish stockrooms), to shoe manufacturing 
(ordering findings and other purchased supplies). In particular, the ap- 
proach has been found helpful in controlling stocks made up of many 
low-value items used regularly in large quantities. 


Qll — 


Guides to Inventory Policy I 229 

A number of realistic complications might have been introduced into 
the. Brown and Brown, Inc., problem. For example: 

In determining the size of a manufacturing run, it sometimes is important 
to account explicitly for the production and sales rate. In this case, the inven- 
tory balance pattern looks like exhibit iv instead of the saw-tooth design in 
exhibit 1. The maximum inventory point is not equal to the amount produced 
in an individual run, but to that quantity less the amount sold during the 
course of the run. The maximum inventory equals Q (1 — S/P), where Q is 
the amount produced in a single run, and S and P are the daily sales and pro- 
duction rates respectively. 

This refinement can be important, particularly if the sales rate is fairly large 
compared with the production rate. Thus, if the sales rate is half the produc- 
tion rate, then the maximum inventory is only half the quantity made in one 
run, and the average inventory equals only one-fourth the individual run 
quantity. This means that substantially more inventory can be carried— in fact, 
about 40% more. 

When a number of products are made on a regular cycle, one after another, 
with the sequence in the cycle established by economy in change-over cost, 
the total cycle length can be obtained in the same way as described above. 
Of course, it sometimes happens that there is a periodic breach in the cycle, 
either to make an occasional run of a product with very low sales or to allow 
for planned maintenance of equipment; the very simple run-length formulas 
can be adjusted to allow for this. 

Other kinds of costs can also be included, such as different sorts of han- 
dling costs. Or the inventory cost can be defined in such a way as to include 
transportation, obsolescence, or even capital and storage cost as part of the 
unit value of the product against which a charge for capital is made. When 
a charge for capital is included as part of the base value in computing the cost 
of capital, this is equivalent to requiring that capital earnings be compounded; 
this can have an important bearing on decisions connected with very low 
volume items which might be purchased in relatively large, long-lasting 

Complications such as the foregoing, while important in practice, repre- 
sent changes in arithmetic rather than in basic concept. 


When the analytic approach is applied to Brown and Brown's problem 
and similar cases, it reveals certain relationships which are significant and 
useful to executives concerned with inventory management: 

1. The appropriate order quantity and the average inventory maintained 
do not vary directly vuith sales. In fact, both of these quantities vary with the 
square root of sales. This means that with the same ordering and setup cost 
characteristics, the larger the volume of sales of an item, the less inventory per 
unit of sales is required. One of the sources of inefficiency in many inventory 
control systems is the rigid adoption of a rule for ordering or carrying inven- 
tory equivalent to, say, one month's sales. 

230 Inventory and the Administrator 

2. The total cost in the neighborhood of the optimum order quantity is 
relatively insensitive to moderately small changes in the amount ordered. Ex- 
hibit 11 illustrates this proposition. Thus, all that is needed is just to get in the 
"right ball park," and a good answer can be obtained even with fairly crude 
cost data. For example, suppose the company had estimated that its total cost 
of holding 1,000 casings in inventory for a year was $1 when it actually was $2 
(as in our illustration). Working through the same arithmetic, the company 
would have arrived at an optimum order quantity of 70,000 casings instead of 
50,000. Even so, the total cost would have been (using the correct $2 annual 
carrying cost): 

3.6 trips per year @ $10 = $36 

35,000 casings average inventory @ $0,002 = 70 

Total annual cost = $106 

Thus, an error of a factor of 2 in one cost results in only a 6% difference 
in total cost. 

In summary, Brown and Brown's problem, despite its oversimplification, 
provides an introduction to the analytic approach to inventory problems. 

In particular, it illustrates the first essential in such an approach— i.e., de- 
fining an inventory function. In this case the function is to permit pur- 
chase or manufacture in economical order quantities or run lengths; in 
other cases it may be different. The important point is that this basic 
function can be identified wherever it may be found— in manufacturing, 
purchasing, or warehouse operation. 

The only way to cut inventories is to organize operations so that they 
are tied more closely together. For example, a company can cut its raw 
materials inventory by buying in smaller quantities closer to needs, but it 
does so at a cost; this cost results from the increased clerical operations 
needed to tie the purchasing function more closely to manufacturing and 
to keep it more fully informed of manufacturing's plans and operations. 
The right inventory level is reached when the cost of maintaining any 
additional inventory cushion offsets the saving that the additional inven- 
tory earns by permitting the plant to operate in a somewhat less fully 
organized fashion. 

B. and B.'s problem also illustrates problems and questions connected 
with defining and making costs explicit. The inventory capital cost is 
usually not found on a company's books, but it is implied in some of the 
disagreements over inventory policy. Here, again, bringing the matter 
into the open may help each side in a discussion to recognize its own and 
the others' hidden assumptions, and thus more quickly to reach a com- 
mon agreement. 

Guides to Inventory Policy II 231 


John F. Magee* 

The discussion of factors influencing the inventory is considered in 
light of the reorder problems and also production scheduling. Difficulties 
of policy are illustrated in an extended inventory and production sched- 
uling case. 

Marketing and production executives alike have an immediate, vital 
interest in safety stocks. In these days of strong but often unpredictable 
sales, safety stocks afford, for the factory as well as for the sales office, a 
method of buying short-term protection against the uncertainties of cus- 
tomer demand. They are the additional inventory on hand which can be 
drawn upon in case of emergency during the period between placement 
of an order by the customer and receipt of the material to fill the order. 
However, in practice their potentials are often needlessly lost. 

Our studies have shown that the methods used by existing systems in 
industry often violate sound control concepts. The economy of the com- 
pany is maintained, in the face of instability and inefficiency in the in- 
ventory control system, only because of constant attention, exercise of 
overriding common sense, and use of expediting and other emergency 
measures outside the routine of the system. 

Actually, it is possible to have inventory controls which are not only 
flexible but also carefully designed and explicit. But the task needs special 
analytical tools; in a complicated business it defies common sense judg- 
ment and simple arithmetic. Methods must be employed to take direct 
account of uncertainty and to measure the response characteristics of the 
system and relate them to costs. Such methods are the distinctive mark of 
a really modern, progressive inventory control system. 


Like transit stocks and lot-size stocks, safety stocks "decouple" one 
stage in production and distribution from the next, reducing the amount 
of over-all organization and control needed. 

*From Harvard Business Review, XXXIV, 2 (1956), 103-16. Reprinted by permis- 
sion of the Harvard Business Review. 

232 Inventory and the Administrator 

But the economies of safety inventories are not fairly certain and imme- 
diate. The objective is to arrive at a reasonable balance between the costs 
of the stock and the protection obtained against inventory exhaustion. 
Since exhaustion becomes less likely as the safety inventory increases, 
each additional amount of safety inventory characteristically buys rela- 
tively less protection. The return from increasing inventory balances 
therefore diminishes rapidly. So the question is: How much additional in- 
ventory as safety stock can be economically justified? 

To answer this question we need to look at the two basic systems of in- 
ventory replenishment to handle uncertainty about sales and see how they 
produce different results. 


Under any fixed order system— the old-fashioned "two-bin" system or 
one of its modern varieties— the same quantity of material is always 
ordered (a binful in the primitive system), but the time an order is 
placed is allowed to vary with fluctuations in usage (when the bottom of 
one bin is reached). The objective is to place an order whenever the 
amount on hand is just sufficient to meet a "reasonable" maximum demand 
over the course of the lead time which must be allowed between place- 
ment of the replenishment order and receipt of the material. 

Where the replenishment lead time is long (e.g., three months) com- 
pared with the amount purchased at each order (e.g., a one-month sup- 
ply), there are presumably some purchase orders outstanding all the time 
which, on being filled, will help replenish the existing inventory on hand. 
In such cases, of course, the safety stocks and reorder points should be 
based upon both amount on hand and on order. Where, on the other 
hand, the lead time is short compared with the quantity ordered, as in 
most factory two-bin systems, the amount on hand and the total on hand 
and on order are in fact equivalent at the time of reordering. 

The key to setting the safety stock is the "reasonable" maximum usage 
during the lead time. What is "reasonable" depends partly, of course, on 
the nature of short-term fluctuations in the rate of sale. It also depends— 
and here is where the top executive comes foremost into the picture— on the 
risk that management is prepared to face in running out of stock. What is 
the level of sales or usage beyond which management is prepared to face 
the shortages? For example: 

In exhibit 1, continuing the hypothetical case of Brown and Brown, Inc., 1 
the curve shows the number of weeks in which the demand for casings may be 
expected to equal or exceed any specified level. (Such a curve could be 
roughly plotted according to actual experience modified by such expectations 
or projections as seem warranted; refinement can be added by the use of math- 
ematical analysis when such precision seems desirable.) 

1 See page 225 (Article 24). 

Guides to Inventory Policy II 


Now, if it takes B. and B. a week to replenish its stocks and the manage- 
ment wishes to keep the risk of running out of stock at a point where it will 
be out of stock only once every 20 weeks, or 5% of the time, then it will 
have to schedule the stock replenishment when the inventory of casings on 
hand drops to 66,000 units. Since the expected or average weekly usage is 
50,000 units, the safety stock to be maintained is 16,000 (making a total stock 
of 66,000). 


Percentage of time demand exceeds Level D 

10 20 30 40 50 60 70 
Level of demand (thousands of units) 

This example, of course, assumes a single, rather arbitrary definition of 
what is meant by risk or minimum acceptable level of customer service. 
There are a number of ways of defining the level of service, each ap- 
propriate to particular circumstances. One might be the total volume of 
material or orders delayed; another, the number of customers delayed 
(perhaps only in the case of customers with orders exceeding a certain 
size level), still another the length of the delays. All of these definitions 
are closely related to the "probability distribution" of sales-i.e., to the 
expected pattern of sales in relation to the average. 

Cost of Service Failure. It is easy enough to understand the principle 
that setting a safety stock implies some kind of a management decision or 
judgment with respect to the maximum sales level to be allowed for, or 

2 34 

Inventory and the Administrator 


Percentage of orders delayed 


$80,000 $120,000 

20% annual charge for investment 




the cost of service failure. But here is the rub: service failure cost, though 
real, is far from explicit. It rarely, if ever, appears on the accounting 
records of the company except as it is hidden in extra sales or manufactur- 
ing costs, and it is characteristically very hard to define. What is new in 
inventory control is not an accounting technique for measuring service 
cost but a method of self-examination by management of the intuitive 
assumptions it is making. The progressive company looks at what it is in 
fact assuming as a service-failure cost in order to determine whether the 
assumed figure is anywhere near realistic. 

For example, characteristically one hears the policy flatly stated: "Back 
orders are intolerable." What needs to be done is to convert this absolute, 
qualitative statement into a quantitative one of the type shown in exhibit 
ii. Here we see the facts which might be displayed for the management of 
a hypothetical company to help it decide on a customer service policy: 

Guides to Inventory Policy II 235 

To get a 90% level of customer service (i.e., to fill 90% of the orders im- 
mediately), a little over three weeks' stock must be carried— an investment 
of $64,000 with an annual carrying cost of $12,800. 

Filling another 5% of orders immediately, thereby increasing the service 
level to 95%, would mean about one week's more stock, with an extra annual 
cost of $3,800. 

Filling another 4% immediately ( a 99% service level) would cost an extra 
$7,400 per year. 

At each point the management can decide whether the extra cost is 
justified by the improved service. Thus, the chart becomes a device for 
comparing policies on service and inventories for consistency and ration- 


The periodic reordering system of inventory replenishment— the other 
basic approach to handling uncertainty about demand— is very popular, 
particularly where some type of book inventory control is employed and 
where it is convenient to examine inventory stocks on a definite schedule. 
The idea underlying all varieties of this system is to look at stocks at fixed 
time intervals, and to vary the order amount according to the usage since 
the last review. 

The problem is that many seemingly similar ways of handling a cyclical 
ordering system may have hidden traps. A typical difficulty is instability 
in reordering habits and inventory levels caused by "overcompensation"; 
that is, by attempting to outguess the market and assuming that high or 
low sales at one point, actually due to random causes, indicate an estab- 
lished trend which must be anticipated. 

The most efficient and stable reorder scheme or rule has a very simple 

A forecast or estimate of the amount to be used in the future is made for a 
period equal to the delivery lead time plus one reorder cycle. Then an order 
is placed to bring the total inventory on hand and on order up to the total 
of the amount forecast for the delivery lead and cycle times, plus a standard 
allowance for safety stock. Under such a scheme, the average inventory ex- 
pected to be on hand will be the safety balance plus one-half the expected us- 
age during a reorder cycle. 

Many companies subscribe to this plan wholeheartedly in principle but 
only halfheartedly in practice. A common tendency, for instance, is to 
make the forecast but then, if sales increase, to revise it upward and trans- 
mit the increase back to the plant. The whole value of a safety stock based 
on a balancing of the costs of running out and the costs of rush orders to 
production is thus lost. 

2^6 Inventory and the Administrator 

Readers may recognize the application here of servo theory, the body 
of concepts (including feedback, lags or reaction times, type of control, 
and the notion of stability) developed originally by electrical engineers 
in designing automatic or remotely controlled systems. An inventory sys- 
tem, though not a mechanical device, is a control system and as a conse- 
quence is subject to the same kinds of effects as mechanical control sys- 
tems and can be analyzed using the same basic concepts. 


Each system of reordering inventories has its own advantages. Here are 
the conditions under which the fixed order system is advantageous: 

Where some type of continuous monitoring of the inventory is possible, 
either because the physical stock is seen and readily checked when an item 
is used or because a perpetual inventory record of some type is maintained. 

Where the inventory consists of items of low unit value purchased infre- 
quently in large quantities compared with usage rates; or where otherwise 
there is less need for tight control. 

Where the stock is purchased from an outside supplier and represents a 
minor part of the supplier's total output, or is otherwise obtained from a source 
whose schedule is not tightly linked to the particular item or inventory in 
question; and where irregular orders for the item from the supplier will not 
cause production difficulties. 

For example, the fixed order system is suitable for floor stocks at the 
factory, where a large supply of inexpensive parts (e.g., nuts and bolts) 
can be put out for production workers to draw on without requisitions, 
and where a replenishment is purchased whenever the floor indicates the 
supply on hand has hit the reorder point. 

By contrast, the periodic reordering system is useful under these condi- 

Where tighter and more frequent control is needed because of the value of 
the items. 

Where a large number of items are to be ordered jointly, as in the case of a 
warehouse ordering many items from one factory. (Individual items may be 
shipped in smaller lots, but the freight advantages on large total shipments can 
still be obtained. ) 

Where items representing an important portion of the supplying plant's out- 
put are regularly reordered. 

In general, since safety stocks needed vary directly with the length of 
the period between orders, the periodic system is less well suited where the 
cost of ordering and the low unit value of the item mean infrequent large 

It should be noted that modifications of the simplest fixed order system 
or intermediates between the fixed order system and the periodic reorder- 
ing system are also possible and very often useful; they can combine the 

Guides to Inventory Policy II 237 

better control and cost features of each of the "pure" schemes. For ex- 

One type of scheme often useful— the "base stock" system— is to review in- 
ventory stocks on a periodic basis but to replenish these stocks only when 
stocks on hand and on order have fallen to or below some specified level. 
When this happens, an order is placed to bring the amount on hand and on 
order up to a specified maximum level. 

The choice of frequency of review and the minimum and maximum inven- 
tory points can be determined by analysis similar to that used for the other 
systems, but precautions must be taken— such as that stocks on order must al- 
ways be counted when reorder quantities are figured— in order to avoid prob- 
lems of instability and oscillation which can easily creep into rules that are 
apparently sound and sensible. 

Interaction Among Factors. As mathematical analysis will indicate, the 
safety stock, reorder quantity, and reorder level are not entirely inde- 
pendent under either the fixed order or the periodic reordering system 
(or any combination thereof ) : 

Where the order amount is fixed, the safety stock is protection against 
uncertainty over the replenishment time (measured by the reorder level). But 
it is the size of the order amount that determines the frequency of exposure 
to risk. With a given safety level, the bigger the order placed, the less fre- 
quently will the inventory be exposed to the possibility of run-out and the 
higher will be the level of service. 

Where inventories are reordered on a periodic time cycle, the uncertainty 
against which safety stocks protect extends over the total of the reorder period 
and replenishment time. But here it is the length of the reordering cycle that 
determines the risk. The shorter the period and the closer together the re- 
orders, the less will be the chance of large inventory fluctuations and, as a 
consequence, the less will be the size of safety stock required in order to main- 
tain a given level of service. 

The interaction among the frequency of reorder, the size of reorder, 
and safety stocks is often ignored as being unimportant, even in setting up 
fairly sophisticated inventory control schemes (although the same com- 
panies readily consider the lot-size problem in relation to the other fac- 
tors). In many cases this may be justifiable for the purpose of simplifying 
inventory control, particularly methods for adjusting reorder quantities 
and safety stocks to changing costs and sales. On the other hand, cases do 
arise from time to time where explicit account must be taken of such inter- 
actions so that an efficient system may be developed. 

Note, too, that the factors governing the choice of any reorder scheme 
are always changing. Therefore, management should provide for routine 
review of the costs of the system being used, once a year or oftener, so 
that trends can be quickly identified. Also, control chart procedures, like 
simple quality control methods, should be used to spot "significant" shifts 
in usage rates and in the characteristics of customer demand (fluctuations, 

2 ^8 Inventory and the Administrator 

order size, frequency of order, etc.). Schemes for checking such matters 
each time a reorder point is crossed are easily incorporated in the programs 
of automatic data-handling systems used for inventory control; they can 
also be applied to manual systems, but less easily and hence with some 
temptation to oversimplify them dangerously. 


Now let us turn to the important relationships between safety stocks 
and production. The safety stock affects, and is affected by, production 
run cycles, production "reaction times," and manufacturing capacity 


In production cycling problems, as in periodic reordering, the longer 
the run on each product, the longer one must wait for a rerun of that 
product; therefore, a larger safety stock must be maintained as protection. 
Shorter, more frequent runs give greater flexibility and shorter waiting 
periods between runs, and thus lower safety inventory requirements. Also, 
again the interaction between factors must be taken into account. For 

A chemical company arrived at production run cycles for a set of five prod- 
ucts going through the same equipment on the basis of only setup costs and 
cycle inventories (e.g., lot-size inventories), ignoring the interaction between 
cycle length and safety stocks. It found that on this basis an over-all product 
cycle of approximately 20 days, or one production month, appeared optimum, 
allowing 4 days per product on the average. However, when the problem was 
later re-examined, it was discovered that the uncertainty introduced by long 
lead times was so great that the over-all product cycle could in fact be eco- 
nomically cut back to less than 10 days. Doubling setup costs would be more 
than offset by savings in inventory and storage costs resulting from a reduc- 
tion in the needed safety stocks. 

exhibit in illustrates the cost characteristics found to exist. The three dashed 
lines show separately the annual costs of changeovers, carrying cycle inven- 
tories, and carrying safety stocks, compared with the length of the individual 
production cycle. Adding together only the first two costs leads to the lower 
of the solid lines. This is at a minimum when the production cycle is 20 days 
long, indicating an apparent annual cost of $40,000. However, if all costs are 
included (the solid line at the top), the total annual cost on a 20-day cycle is 
$95,000. On this basis total costs are at a minimum when the cycle is 10 days 
long-only $70,000. This means a saving of $25,000 annually on the products in 

Guides to Inventory Policy II 



Annual cost 

Total cost of 
original cycle 


80,000- - 

Total cost 
adjusted cycle 

60,000- - 


annual 40,000 

20,000- - 

^ ^ — """ * Cycle stoc 

stock cost 

5 10 15 20 

Length of production cycle (days) 



Safety stocks give only short-term protection against sales uncertainty. 
If stocks are being replenished from production, the effectiveness of 
over-all control depends also on the ability to restore them in case of 

If total demand varies, the ability to restore stocks depends, in turn, on 
the ability of the production facilities to react to chance fluctuations. In 
order to get low inventories, the process must have fast reactions properly 
controlled or (equivalently) in some cases large "capacity." If reactions 
are slow or limited, inventories must be large, and the inventory in effect 
serves another type of protective function, namely, protection of produc- 
tion rate or capacity from the stresses of demand fluctuation. 

How fast should production operations respond to sales fluctuations, 
and to what extent should these fluctuations be absorbed by means of in- 
ventory? The costs of warehousing and cash investment in inventory need 
to be balanced against the costs of changing production rates or building 
excess capacity into the production system. 

The actual cost of making out schedules, which depends on the fre- 
quency with which they are made and the degree of precision required, 
also should be considered, as well as the speed of reaction of production 
which is physically possible (e.g., the employee training time). When 

240 Inventory and the Administrator 

these costs are made explicit, management may find itself having to bal- 
ance conflicting objectives. To illustrate: 

A metal fabricator making a wide line of products to order attempted to 
provide immediate service to customers. He found that on the average his de- 
partments needed a substantial excess of labor over the normal requirements 
of the jobs flowing through, and this excess was essentially idle time. On the 
other hand, when he attempted to cut the excess too thin, backlogs began to 
build up. He had to weigh his desire to get the lead time down against the 
costs of excess unused labor. 

Ordinarily we want to avoid passing back the full period-to-period sales 
fluctuation by making corresponding changes in the size of orders placed 
on production because it is uneconomical. What we can do instead is to: 

1. Set the production level in each period equal to anticipated needs over 
the lead time plus the scheduling period not already scheduled, plus or minus 
some fraction of the difference between desired and actual inventory on hand. 

2. Alternatively, change the existing production level or rate by some frac- 
tion of the difference between the existing rate and the rate suggested by the 
simple reorder rule (i.e., that an order be placed in each period equal to the 
anticipated requirements over the lead time plus the scheduling period, plus 
or minus the difference between desired and actual inventory on hand and 
on order). 

Each of these alternatives is useful in certain types of plants, depending 
on whether the cost of production fluctuations comes primarily from, 
say, overtime and undertime (work guarantee) costs or from hiring, train- 
ing, and layoff costs. Each in appropriate circumstances will lead to 
smoother production, at the expense of extra inventory to maintain the 
desired level of service. 

When the different costs involved are identified and measured, mathe- 
matical techniques can be used to show the effect that varying the num- 
bers in the rule (in particular, the size of the fraction used) has on in- 
ventory and production expense and to arrive at an economical balance 
between the needs of marketing and manufacturing. These two rules are 
expressions of servo theory, like that referred to earlier in connection 
with inventory. Here it may be worthwhile to see in some working detail 
how the theory can be applied mathematically: 

The first rule can be stated as follows: 

T T 

P« = 2 F i+k - 2 Pi-* + k(h - h) ; k < 1 

fc = k = 1 

Pi is the amount scheduled for production in period i, Fi is the forecast 
requirements for period i, L is the desired inventory, Ii is the actual opening 
inventory on hand in period i, and k is the response number which indicates 
what fraction of the inventory error or production rate departure is to be ac- 
counted for each period. 

Guides to Inventory Policy II 


The fluctuations in inventory resulting from a choice of k in the first rule 
can be expressed as a function of the fluctuations in sales about the forecast, as 
follows (if fluctuations from month to month are not correlated): 


T(ik-k 2 ) + 1 
ik -k 2 

where oj\ is the standard deviation of inventory levels, and o> is the standard 
deviation of actual sales about forecast sales each period. Similarly, the produc- 
tion rate variations resulting from any choice of k can be expressed as: 


The influence of the choice of a response number, k, on the standard devia- 
tion of inventories and on the standard deviation of production rates under 
the first type of rule is shown in exhibit iv. Frequently the costs of produc- 
tion fluctuations are more or less directly proportional to the standard devia- 
tion of fluctuations in the production rate, a measure of the amount of change 
in production level which can be expected to occur. On the other hand, the 
normal inventory level, the average level expected, must be set large enough 
so that even with expected inventory fluctuations, service failures will not 
occur excessively. This means that the larger the standard deviation in inven- 
tory levels, the larger must be the normal level, generally in proportion. There- 
fore, one can "buy" production flexibility with larger inventories, and vice 
versa, with the particular costs in the process concerned determining the eco- 
nomical balance. 





2.0 • 





Response number k 

242 Inventory and the Administrator 

The second rule can be worked through similarly. Here P* is the changed 
amount scheduled for production, and the rule can be stated as follows: 

P*i=P*i- 1 +k(P i -P* i _ 1 );k^I =(j - k)P* i _ 1 +kPi 


T T 

Pi = 2 Fi ♦ * - 2 P*« - * + (/« - L) 

k = & = 1 


In some cases— particularly where output cannot be stocked easily— the 
problem of controlling the production level is not so much one of adjust- 
ing the level to respond to fluctuations in demand, as of setting the ca- 
pacity of the plant or operation at a high enough level to permit demand 
fluctuations to be absorbed without excessive delay. If the capacity is set 
equal only to the desired average rate, fluctuations in demand about this 
desired rate must either be absorbed by inventories or by orders piling 
up in a back-log. 

A theory of such processes is growing; it is known as waiting-line 
theory. This is really a branch of probability theory, and is itself a whole 
body of mathematical techniqu6s and explicit concepts providing a mathe- 
matical framework within which waiting-line and similar problems can be 

Some examples of applications in production scheduling are: flow of 
orders through departments in a job shop; flow of items through the 
stages in an assembly line; clerical processing of orders for manufacture 
or shipping; filling orders in a warehouse or stockroom; and setting up 
shipping or berth facilities to handle trucks or other transport units. In 
each case, fairly well-fixed crews or facilities have to be set up for han- 
dling fluctuating orders or items quickly, avoiding delays in service. A 
balance between the cost of extra personnel or facilities and delays in tak- 
ing care of demand is needed. 

In applying waiting-line theory to such problems, the flow of orders or 
demand for goods can be considered as a demand for service, analogous 
to subscriber cost in a telephone exchange. Orders are handled by one or 
more processing stations, analogous to telephone trunk lines. When the 
order or unit is produced, the processing station is free to take on the next 
order in line, as when a call is completed through the exchange. 


The choice and use of appropriate techniques for inventory control is 
not a simple matter. It takes a good deal of research into sales and product 

Guides to Inventory Policy II 243 

characteristics, plus skill in sensing which of many possible approaches 
are likely to be fruitful. 

To describe these techniques, I shall take a case illustration. This case is 
drawn from a great deal of business experience, but in order to keep the 
detail and arithmetic within manageable proportions without distorting 
the essential points, I have simplified and combined everything into one 
fictional situation. 

Any of the stages of the company's progress toward more efficient in- 
ventory management-from the original to the final-might be found to 
exist in the inventory control practices of a number of sizable companies 
with reputations for progressive and efficient management. These stages 
of advancement in the refinement of inventory control should not be used 
to compare the inventory system of one company or division with that 
of another, for the reasons just mentioned; but they may prove helpful 
to management in answering the questions, "Where are we now?" and 
"What could we do better?" 

Briefly, the case situation is as follows: 

One division of the Hibernian Bay Company makes and sells a small machine 
part. Sales run slightly over 5,000 units annually, and the price is $100 apiece. 
Customers are supplied from four branch stock points scattered about the 
country, which in turn are supplied by the factory warehouse. The machining 
and assembly operations are conducted in a small plant, employing largely 
semiskilled female help. The level of production can be changed fairly rapidly 
but at the cost of training or retraining workers, personnel office expenses, and 
increased inspection and quality problems. The division management has al- 
most complete autonomy over its operations, although its profit records are 
closely scrutinized at headquarters in Chicago. 

Originally the factory and branch warehouse stocking practices were hap- 
hazard and unsatisfactory. In total, nearly four months' stock was carried in 
branches, in the factory warehouse, or in incompleted production orders. A 
stock clerk in each branch who watched inventories and placed reorders on 
the factory warehouse was under pressure to be sure that stocks were adequate 
to fill customer orders. The factory warehouse reorder clerk in turn watched 
factory stocks and placed production orders. Production runs or batches were 
each put through the plant as a unit. Fluctuations in production, even with 
apparently sizable stocks on hand, caused the management deep concern. 


The management decided to try to improve inventory practices and 
appointed a research team to study the problem. The team suggested using 
"economical order quantities" for branch orders on the factory warehouse 
and warehouse orders on production, as a basis for better control. The 
steps followed were: 

The research team suggested that the formula for determining the economi- 
cal order quantity was x = ViAs/i, where A = fixed cost connected with an 
order (setup of machines, writing order, checking receipts, etc.), i— annual 


Inventory and the Administrator 


A. Presumed Operation 





N. Order 
\^ placed 

Order placed 



. Reorder X^ 
Point X 


Safety stock 


1 1 

1 1 

1 1 

12 3 4 5 

Time (weeks) 


B. Actual Operation 

-l 1 r 

12 3 4 5 

Time (weeks) 

Guides to Inventory Policy II 245 

cost of carrying a unit in inventory, s = annual movement, and x — "eco- 
nomical order quantity." 

The team found that each branch sold an average of 25 units a week, or 
1,300 per year; that the cost of a branch's placing and receiving an order was 
$19 ($6 in clerical costs at the branch and factory, $13 in costs of packing and 
shipping goods, receiving, and stocking); that annual inventory carrying costs 
in the branches were $5 per unit, based on a desired 10% return on incre- 
mental inventory investment. The reorder quantity for each branch was com- 
puted as V 2 ' $ 10 ' M°°/$5 = I0 ° unit reo r cl er quantity. 

A system was set up where each branch ordered in quantities of 100, on the 
average, every four weeks. On this basis, without further action, each branch 
would have had an average inventory of one-half a reorder quantity, or 50 
units. (The books would show 75 units, since stock in transit from factory 
warehouse to branch was also charged to the branch, and with average transit 
time of one week this would average 25 units.) 

The next step was to provide for enough to be on hand when a reorder 
was placed to last until the order was received. While the average transit time 
was one week, experience showed that delays at the factory might mean an 
order would not be received at the branch for two weeks. So sales for two 
weeks had to be covered. 

Statistical analysis showed that sales in any one branch over two weeks could 
easily fluctuate from 38 units to 62 units and could conceivably go as high as 
65-70. The management decided that a 1 % chance of a branch running out of 
stock before getting an order would be adequate. 

Calculations then indicated that the maximum reasonable two-week demand 
to provide for would be 67. (The statistical basis was that sales fluctuate about 
the average at random; that fluctuations in the various branches are inde- 
pendent of one another; and that the standard deviation is y/st where s = sales 
rate, and t = length of individual time period.) 

The branches therefore were instructed to order 100 units whenever the 
stock on hand and on order was 67 or less. This gave an inventory in each 
branch made up on the average as follows: 

Safety stock 42 (order point, 67, less normal week's usage, 25) 

Order cycle stock 50 (one half 100-unit order) 

In transit 25 (one week's sales) 

Total 117 or 4.7 weeks' sales 

The resulting behavior of the reorder system is shown in exhibit v- 
both as it would be presumed in theory and as it actually turned out. Al- 
though the actual performance was much less regular than presumed, the 
two compare fairly well— testimony to the soundness of the procedure. 


At the factory warehouse end, the "economical order quantity" scheme 
worked as follows: 

The cost of holding a unit in inventory was $3.50 per year (at 10% return 
on investment); the cost of placing an order and setting up equipment for each 
order was $13.50; and, of course, a total of 5,200 units was made each year. 

246 Inventory and the Administrator 

These indicated that each production order should be for V 2 ' $13-50 ■ 5,200/ 
$3.50 = 200 units. 

Factory processing time was two weeks; it would take two weeks for each 
order to reach the warehouse. The warehouse would need to place its re- 
plenishment order on the factory when it had enough on hand or on order 
to fill maximum reasonable demand during the next two weeks. 

On the average, the factory warehouse would receive one order a week from 
the branches (one every four weeks from each of four branches) under the 
new branch reorder system. In fact, because of the fluctuations in branch sales 
described before, it was found that orders on the factory warehouse fluctu- 
ated substantially in any two-week period (see exhibit vi). 


Number of 

Number of items 


branch orders 


of weeks 

A. Weekly Periods 












B. Biweekly Periods 





















7 + 



It was agreed that to give branches service adequate to maintain their 
own service, stocks at the factory warehouses would have to be high 
enough to fill demand 99% of the time, i.e., a replenishment order would 
have to be placed when 600 units were on hand. This meant a safety stock 
of 600 units minus 200 (normal usage), or 400 units. Cycle stock averaged 
half a run, or 100 units, and stock in process an additional half run, or 100 
units. Total factory stock, then, was: 

Cycle stock 100 units 

Stock in process 100 

Safety stock 400 

Total 600 units 

exhibit vii gives a picture of the apparent costs of the "economical 
order" system. The stock of 1,068 units equaled less than 11 weeks' sales, 
a fairly substantial reduction, and the management felt that it had a better 
control, since clerical procedures were set up to adapt readily to any 

Guides to Inventory Policy II 


changes in inventory charges (currently 10% per year) or service level 
requirements the management might choose to make. 



Cost each 

Annual cost 

4 branches 

600 units 
468 units 



Reorder cost 


5 2 /year 






But the factory still had problems. On the average, the warehouse 
would place one production order every two weeks, but experience 
showed that in 60% of the weeks no orders were placed, in 30% one 
order, and in 10% two, three, or more orders were placed, exhibit viii 
shows orders on the factory and the production level for a representative 
period of weeks. 


Orders placed 

Production level (units/week) 





□ Production level 
g2 Orders placed 






8 12 16 20 24 
Time (weeks) 



28 32 

248 Inventory and the Administrator 

Factory snarls due to these fluctuations occasionally caused the factory 
to miss deadlines. These in turn led on occasion to warehouse delays in 
filling branch orders, and forced the branches to hold to the two-week 
delivery time even though actual transit time was only one week. An an- 
alysis revealed the following: 

Factory fluctuations were very costly. A statistical regression of costs against 
operating levels and changes showed that annual production costs were af- 
fected more by the average size of changes in level than by the frequency of 
change; a few large changes in operating level were much more costly than 
many small changes. 

Under the "economical reorder quantity" system, production fluctuations 
were no larger than before, but the average change up or down actually 
equaled 80% of the average production level. This was estimated to cost 
$11,500 annually, bringing the total cost of the system, including costs of 
holding inventories, placing orders, and changing production rates, to 
$17,280 per year. 

This led to the suggestion that the company try a new scheme so that 
orders on the factory warehouse and the factory would be more regular. 
A system with a fixed reorder cycle or period was devised, under which 
branch warehouses would place orders at fixed intervals, the order being 
for the amount sold in the period just ended. The factory warehouse 
would ship the replenishment supply, order an equivalent amount from 
the factory, and receive the order within two weeks or by the beginning 
of the next review period, whichever was longer. 

Under this scheme, each branch warehouse would need to keep its 
stock on hand or on order sufficient to fill maximum reasonable demand 
during one review period plus delivery time (tentatively taken as two 
weeks) on the basis of the reorder rule described previously in this article. 
The question to be determined was: How long should the review period, 
that is, the time between reorders, be? exhibit ix summarizes inventories 
and costs for reorder intervals ranging from one to six weeks, based on the 
following facts and figures: 

1. Branch safety stock was determined from a study of branch sales fluc- 
tuations, to allow for maximum reasonable demand over the reorder interval 
plus the two-week delivery period. 

"Maximum reasonable demand" was defined to allow a 0.25% risk of being 
out of stock in any one week (equal to the 1% risk on the average four-week 
interval under the "economical reorder quantity" system described previously). 

2. Branch cycle stock would average one-half of an average shipment. 
Under this system, the average shipment to a branch each period would equal 
the average sales by the branch in one period (25 units X number of weeks). 

Guides to Inventory Policy II 



Length of period (weeks) 







Branch warehouse 

Safety stock 
Cycle stock 
Transit stock 









Total units of stock 





1 17-5 

1 31.0 

Annual inventory cost 
Ordering cost 

$ 310 

$ 380 

$ 450 

$ 5i5 

$ 590 

$ 650 

Total cost each branch 


% 875 

$ 780 

% 765 

% 785 

$ 815 

Total cost four branches 







Factory warehouse 
Safety stock 
Cycle stock 










Total units of stock 




Annual inventory cost 
Ordering cost 

$ 290 

$ 990 

$ 465 

$ 815 


$ 670 

$ 905 


$ 865 






Total cost factory 
Production change costs 


Total system costs 







3. Transit stock equaled one week's sales. 

4. Branch inventory carrying cost was $5 per unit per year. 

5. Branch ordering costs equaled $19 per order, with one order per period. 
A one-week period would mean 52 orders per year; a two-week period, 26 
orders per year; etc. 

6. Factory safety stock was set to allow a 1% risk that the warehouse 
would be unable to replenish all branch shipments immediately. 

7. Factory cycle stock in process or in the warehouse would be approxi- 
mately equal to one-half the sales in any one period. 

8. Factory inventory carrying cost was $3.50 per unit per year. 

9. Factory ordering costs equaled $13.50 per order (see 5 above). 

10. Production change costs were proportional to the period-to-period 
changes in production level, equal under this system to period-to-period 
changes in branch sales. 

The figures show that a two-week reorder interval would be most eco- 
nomical for the company as a whole, and this was chosen. Costs were 
estimated to be $6,600, compared with $17,300 under the "economical 
reorder quantity" system. While the new system cut total inventories by 
nearly 70%, most of the gain came from smoother production operations. 
exhibit x shows weekly production for a representative period under the 
new system. 

2 5 

Inventory and the Administrator 


Production level (units/week) 






_H r. 




50 j_ 

T T 

8 12 16 20 24 28 32 
Time (weeks) 

Further economies became apparent when the system was in operation: 

i. The reduction in production fluctuations made it possible to meet pro- 
duction deadlines regularly, cutting the effective lead time in deliveries to 
branches and thereby permitting modest reductions in branch safety stocks. 

2. The inventory system was found well suited to "open" production 
orders. Instead of issuing a new order with each run, the moderate fluctuations 
made it possible to replace production orders with simplified "adjusting 
memos" and at the same time to eliminate much of the machine setups. 


The success with the periodic reordering system encouraged the com- 
pany to go further and try the "base stock" system referred to earlier. 
Under this system, the branch warehouses would report sales periodically. 
The factory would consolidate these and put an equivalent amount into 
production. Stocks at any branch would be replenished whenever re- 
ported sales totaled an economical shipping quantity. 

Two possible advantages of this system compared to the fixed period 
scheme were: (i) Branches might be able to justify weekly sales reports, 
reducing production fluctuations and safety stock needs still further. (2) 
It might be possible to make less frequent shipments from factory to 
branches and make further savings. The following questions had to be 

How frequently should branches report sales? As noted earlier, cost studies 
showed that of the $19 total cost of ordering and receiving goods $6 
represented clerical costs in placing and recording the order. Here is a sum- 
mary of the costs affected by the choice of reporting interval: 

Guides to Inventory Policy II 


Branch safety stock 
Production changes 
Branch clerical costs 


Reporting interval 

One week 



$ 500 

4 X 52 1,250 


Two weeks 
Number Cost 

108 % 540 


4 X 26 625 


Thus, there appeared to be some advantage to reporting sales weekly from 
branches to the factory. 

How big should replenishment shipments be? exhibit xi summarizes the 
system costs related to the size of shipment from factory to branch. Each line 
shows the total of the cost indicated plus those represented by the line below. 
The total system cost (top line) is lowest at 82; that point is therefore the op- 
timum shipping quantity from factory to branch warehouse. The same answer 
can be obtained from the formula given before, V 2 ' $ J 3 " 1,3 00/$ 5 = 82. 

The base stock system therefore was set up with weekly reporting and re- 
plenishment shipments of 82 units to branches. The total cost of the base stock 
system was $5,200 compared with $6,600 under the previous system. 


The company, cheered by its successes, decided to see if even further 
improvements might be obtained by cutting down further on production 










Total system cost 


25 50 75 100 125 150 

Size of shipment (units) 

Shipping cost 

Inventory cost related 
to order quantity 

Process inventory cost 
Reporting cost 

Production changes cost 
Safety stock 

5 : 

Inventory and the Administrator 

fluctuations. As it was, the production level under the base stock system 
was being adjusted each week to account for the full excess or deficiency 
in inventory due to sales fluctuations. It was proposed that production be 
adjusted to take up only a fraction of the difference between actual and 
desired stocks, with added inventories used to make up the difference. 

The possibilities were analyzed along the lines described previously in the 
text: the results are summarized in exhibit xii. The two costs that would 



$ 500 

production changes 


8 1.0 

be affected are costs of changing production and costs of holding inven- 
tories, in particular safety stocks. These are affected by the fraction of 
the inventory departure that is made up each week by adjusting pro- 

The study showed that the cost would be minimized with the rate of 
response set equal to 0.125, as seen in the exhibit. (This compared with a 
response rate of 1.0 under the base stock system.) The additional savings 
of $970 brought the annual cost of the system down to $4,200. 

Guides to Inventory Policy II 



The results of all the changes made by the division management were 

1. A major reduction in stocks— They had been cut 35% from what they 
were even with the "economical reorder quantity" system. 

2. A substantial reduction in production fluctuations— exhibit xiii shows 
what weekly production levels for a typical period looked like at the end, con- 
trasted with exhibits viii and x for the same sales. 

The problems of the case are common even among the best-run busi- 
nesses and can be solved in much the same way with much the same re- 
sults. Of course, a large part of the effort and expense that were necessary 
in this step-by-step, evolutionary approach could be saved. Technical 
methods are available for analyzing and measuring the performance of 
alternate systems so that management can proceed directly to the ultimate 


Production level (units/week) 



12 16 20 
Time (weeks) 

24 28 32 

system that is most desirable; management does not have to feel its way. 
Let me emphasize again, however, that no one kind of system should be 
considered "the goal." The efficiency of any given inventory control plan 
depends too much on the demand and cost characteristics of the business. 
In the discussion thus far, several large questions remain unanswered. 
What happens when the business is subject to seasonal sales? What more 

254 Inventory and the Administrator 

can be done than to insure that desired levels of service are maintained 
while cutting inventory and production costs? Where do forecasting 
and scheduling fit into the picture? I shall discuss these questions in the 
next and final article in this series. 


John F. Magee* 

Mr. John F. Magee concludes the series of articles by discussing the 
crash, seasonal and similar problems as they affect a rational program of 
inventory and production. Plamiing, with its major problem of forecast- 
ing, and control are included in the total approach to the inventory- 
related aspects of business. 

Businessmen are prone to view inventories with distaste, as an ap- 
parently necessary drain on resources, something that no one has been 
able to eliminate but hardly a "productive" asset like a new machine or 
tool. In fact, however, inventories are as productive of earnings as other 
types of capital investment. They serve as the lubrication and springing 
for a production-distribution system, keeping it from burning out or 
breaking down under external shocks. They help to absorb the effects of 
errors in forecasting demand, to permit more effective use of facilities and 
staff in the face of fluctuations in sales, and to isolate one part of the 
system from the next in order to permit each to work more effectively. 

In this article let us look at the function of a third type of inventory, 
one which is of particular importance in long-range planning: anticipation 
stocks. This type of inventory is most commonly needed where sales are 
highly seasonal, and where either one or the other of these problems 

1. The "crash" or short-peak season problem which arises, for example, in 
the toy industry before Christmas or in certain fashion clothing lines at 
various times during the year. 

2. The more conventional seasonal problem arising in industries where 
sales show a pronounced seasonal swing, with the peak season often extend- 
ing over several weeks or months, as in the case of automobiles, many kinds 
of building materials, certain cosmetics, some types of home appliances, agri- 
cultural supplies, and furniture. 

* From Harvard Business Review, XXXIV, 3 (1956), 57-70. Reprinted by permis- 
sion of the Harvard Business Review. 

Guides to Inventory Policy HI 255 

Stocks built up to buffer production against seasonal fluctuations in 
sales ar,e not the only form of anticipation stocks. Anticipation stocks may 
also be carried, for example, to meet a planned intensive sales campaign or 
to carry sales over a plant vacation or maintenance shutdown. However, 
the questions and methods of attack which apply to seasonally fluctuating 
sales also illustrate approaches to control of other types of anticipation 
stocks; I shall therefore use the former as a basis of discussion in this 


In the "crash" type of problem, management must balance the risks of 
not having enough stocks to fill demand and thus losing profit, or of being 
forced to go to extraordinary measures to buy or produce to fill demand, 
against the risks of having too much on hand and consequently incurring 
sizable write-off and obsolescence loss or storage expense until the next 
selling season. 

The question boils down to how much stock to have on hand when 
the main selling season opens. The objective basically is to have enough 
on hand so that the company can expect, on the average, to break even 
on the last unit produced; that is, to carry enough so that on the last unit 
the expected risk of loss due to inability to fill demand equals the expected 
cost of carrying the unit through to the next season. 


In principle, the solution to the "crash" problem is quite simple. The 
classic "newsboy" case is as good an illustration as any: 

A newsboy has, on the average, 10 customers a night who are willing to 
buy papers costing 50 each. The newsboy makes a profit of 30 on each 
paper he sells, and loses it on each paper he takes out but fails to sell. Let 
us suppose he has kept records, and that 40% of the time he can sell at least 

10 papers and 20% of the time he can sell at least 12 papers. 

If the newsboy does not know how many papers he will actually sell in 
any given day but every days takes out 10 papers, he has a 40% chance of 
selling all the papers and making 30 each, and a 60% chance of not selling 
all papers and losing i0 on each not sold. He can expect the tenth 
paper to produce, on the average over time, a profit of o.60 (30 X 40% 
— 10 X 60%). On the other hand, if he takes 12 papers every night, he can 
expect the twelfth paper to produce, on the average over time, a loss of o.20 
(30 X 20% - 10 X 80%). 

It would not, therefore, be worth his while to take out 12 papers. As a 
matter of fact, he would probably make the greatest total profit by taking 

11 regularly, since he could expect, on the average over time, to do slightly 
better than break even on the eleventh paper (30 X 30% — i0 X 70%, 
or o.20). (See also p. 294.) 

256 Inventory and the Administrator 

The newsboy problem is, after all, not so different from many business 
problems. Certainly from the newsboy's point of view the papers he buys 
which he may not sell represent a lot of money and a sizable risk of his 
capital. Indeed, perhaps the most important difference between the news- 
boy and businessmen in other situations is that the newsboy has to make 
this decision very frequently and therefore has more of a chance to build 
up a lot of experience on which to base intuitive judgments— that is, less 
need for careful calculation or formal statistical methods to wring out of 
past experience the information which is of value. 


Suppose, for example, you are selling cosmetics and you want to make 
up a special Christmas package in a holiday wrapping containing three 
normally separate items at a combined price. You have tried a number of 
deals of this type in the past, and on the whole they have been highly 
profitable. However, individually they have been unpredictable; some 
have been very successful, and some that seemed excellent on paper turned 
out to be failures. 

Your market research manager makes a volume prediction each time; 
on the average, his estimates come fairly close, but rarely on the nose. 
About half the time they are too high and half the time too low. In fact, 
25% of the time your experience shows his estimate to be 20% or more 
on the high side, and just as frequently he misses as badly in the other 
direction. About 10% of the time he is as much as 40% off in each 
direction, and occasionally he really misses and actual sales are 75% or 
more off from the estimate. You are doing everything you can to improve 
these estimates, but in the meantime you have to decide how much to 
make up for your Christmas deal. 

Cost estimates indicate that if a package is not sold, the items can be 
repackaged at an extra cost of about $1 per package. If demand exceeds 
the original run, the extra cost of a special rerun plus emergency ship- 
ments to field stocks is estimated to be $1.75 per package. Following 
reasoning like that in exhibit i (simply a generalized expression in 
mathematical terms of the solution to the newsboy case), you or your 
operations research analyst concludes that you should plan initially to 
have enough stock so that the chance that demand, as it materializes, will 
be covered by the initial run equals the ratio of the special makeup cost 
to the total of (a) special makeup cost plus (b) repackaging loss on un- 
sold items. In other words, you want to make enough so that the chance 
that total sales will be covered by the initial run equals $1.75 -f- ($1.00 -j- 
$1.75), or 64%. With your past experience on forecasting success, this 
means about a 10% overstock; that is, your initial run should exceed your 
estimated needs by about 10%. 

Guides to Inventory Policy III 257 

This will not eliminate all the difficulties by any means. There is nearly 
a 40% chance you will have to make some additional high-cost stock, and 

to "crash" PROBLEM 


V — volume of demand 
f(V) = the probability density function of demand (i.e., distribution of 
demand during one period) 


/ f ( V)dV — the likelihood of selling an amount Y or more during a season 


n — the variable cost of making and holding a unit of stock in in- 
ventory during the selling period, including the capital charge 
for inventory investment, etc. 
m — the profit per unit sold 
L = the cost per unit of not filling an order (loss of good will), over 

and above the loss of profit 
P = the cost of carrying a unit of inventory if unsold by the end of 

the period 
K = the size of the inventory on hand at the beginning of the season 
Then the profit earned during the replenishment cycle is given by: 

p = mV - P(K - V) - nK- V < K 
= mK - L(V -K) -nK-V>K 

and the expected profit earned during the replenishment cycle, E(p), is 
given by: 

E(p) =mfVf(V)dV -nK + mK^j{V)dV - 

Lf (V -K)f(V)dV -Pf(K- V)f(V)dV 

Again, differentiating the expected profit with respect to the inventory on 
hand at the beginning of the season, K, yields: 

dE(p) k 

K J_L - - n + ( m + L) - (M + L + P)ff(V)dV 

dK © 

The maximum profit will be earned when dE(p)/dK = o; that is, when 

k m + L — n 

ff(V)dV = 

o m+L+P 

there is still a good chance you will have unsold goods on hand after the 
holiday. However, this initial decision is about the best you can do with 

258 Inventory and the Administrator 

present forecasting and manufacturing methods to get the right balance 
between the two risks and thus minimize the over-all cost. 

In problems like those -noted above, the costs may, superficially at least, 
look different, and the mathematical details of formulating an approach 
to the problem and arriving at an answer may differ, but the basic elements 
are the same— balancing the costs and lost profit opportunities of demand 
exceeding available stock against the costs and losses of having available 
unused stock or capacity. 


Sometimes from the scanty experience gained in early-season selling 
enough information can be developed so that estimates of total season 
sales and resulting production plans can be adjusted. As more and better 
information becomes available, mathematical methods can be used to alter 
the "strategy" for the season slowly, according to predetermined rules. 

Such a "developing" approach to inventory problems rests on the basic 
premises that one does not know the future, that there is therefore no 
need to plan into it very far in great detail, and that a good strategy for 
the present is one which puts you in a position to make a good choice the 
next time you have a chance, whatever actual experience may develop in 
the meantime. Applications of this general line of approach to problems 
are beginning to be made in the planning of heating-oil production, 
seasonal clothing production, and other seasonal, erratic demand problems. 


In many industries, the basic yearly pattern of seasonal sales may be 
quite predictable, and the over-all volume can be reasonably well esti- 
mated. There may be only a small error of a few percentage points in 
estimating either the total volume or the size of the peak. In situations of 
this sort there are three problems: 

1. Adjusting the forecast of expected sales to allow for safety stocks so as 
to protect against forecast errors. (Examples of an original and an adjusted 
"maximum" sales forecast are shown in exhibit ii. The latter is the original 
cumulative forecast increased by the safety stock allowance. ) 

2. Laying out a production pattern or plan to meet the forecast. (The 
difference between forecast and production plan will result in a planned in- 
ventory as illustrated in exhibit ii. The total costs of inventory and production 
depend on the form of the production curve, and characteristically the object 
is to choose this curve or production plan to minimize the expected total of 
these costs.) 

3. Controlling or adjusting the production plan to keep it aligned with the 
sales forecast, as actual sales experience modifies the forecast and/or results 
in depleted or excessive inventory as compared with the plan. 

Guides to Inventory Policy III 



The answer to the first problem depends somewhat on that for the 
third, as the discussion on production control rules in the second article 
in this series may suggest. In general, however, it is fair to say that in most 
businesses the risks and costs of back orders so outweigh inventory cost 
that substantial protection in the form of safety stocks is justified. These 
safety stocks must be large enough so that stocks can be restored after a 



80,000 - 
70,000 - 



60,000 - CUMULATIVE 


50,000 I- // . 

/ / 
PLANNED / / .* 
40,000 |- INVENTORY / j / 

1 / 
30,000 h v # 

20,000 - 



< 1 1 1 — I !- 



sudden unexpected sales spurt by a smooth and moderate adjustment in 
production rate. The production response rules described in the previous 
article, which take into account the nature of forecast errors, inventory 
costs, and service requirements, are one way of determining what is 
"large enough." 


Inventory and the Administrator 

Another very similar approach is to begin with a forecast of maximum 
expected demand, or maximum demand the company is prepared to plan 
for. The long-range production plan is made out to meet this directly. 
Then production is adjusted downward from plan as excess inventories 
accumulate because of actual sales falling below the maximum plan. (More 
will be said later about this problem of production control.) 


Once the adjusted sales forecast or forecast plus safety stock has been 
obtained, the task is to plan the production rate or draw in the production 
curve shown in exhibit ii. The problem is to find a curve or mathematical 
function that will minimize the total of production and inventory costs. 
In theory, this sounds like a straightforward mathematical problem often 
encountered in physics. In practice, the job is not so easy, but a number 
of techniques have been found useful. 

Graphical Techniques. Where the problem of planning production 
against forecasted seasonal sales is not made too complicated by a variety 
of items, processes, and stages, simple graphical or arithmetic techniques 
can often be useful. For example: 

Suppose a company has a forecast at the beginning of the year which calls 
for requirements as outlined in exhibit hi. The first column shows expected 
sales month by month; the second column shows accumulated expected sales; 
the third column shows a safety reserve to cushion the company against fore- 
cast errors, allowing time for smooth adjustment (the basis for this reserve 
will be discussed further below); and the last column shows the total amount 
that must be produced by the end of each month, allowing for an opening 
stock of 3,500 units. 



(In units) 






total re- 










































1 2,000 
























* Less opening stock of 3,500 

Guides to Inventory Policy III 


The cumulative forecast and cumulative requirements, including opening 
stock, are shown in exhibit iv. The company could produce at an average 
annual rate of 100,000 units, or 8,333 units per month— the production plan 
shown as a straight line in the exhibit. This plan would produce just enough 
inventory at the end-of-year peak to meet requirements. The month-end in- 
ventories (equal to the difference between the production plan and the cumu- 
lative sales forecast) are shown in exhibit v. They average 12,800 units, of 


110,000 - 

100,000 - 
90,000 - 







H 1 1 1 1 1 1 1 1 1 1 1 



which 3,400 are accounted for as safety stock, leaving an average seasonal 
anticipation stock of 9,400 units. If the annual inventory carrying cost were 
$45 per unit, the seasonal anticipation stocks would be costing about $425,000 
per year. 

Various alternatives might be tried to reduce this cost. For example, opera- 
tions might be run during the low months of the year at the rate of 4,000 

262 Inventory and the Administrator 

units per month, building up to a peak rate of over 13,000 units per month in 
September. This plan, shown by the dashed line segments in exhibit iv, would 
result in substantially lower anticipation stocks. The average inventory would 
be 3,700 units, with 3,400 units safety stock, or 300 units seasonal anticipation 
stock. At $45 a unit, the cost of seasonal stock under this plan would be only 


(In units) 


























monthly inventory 


Average safety reserve 


Average seasonal anticipation stock 


$13,000 per year, a saving in inventory cost of over $400,000 per year. 

The saving, of course, is not all net saving, since it is gained at the cost of 
adding and laying off the equivalent of some 9,200 units of production capacity. 
If this were, say, a chemical plant operating well under capacity and the varia- 
tion from 4,000 to 13,200 units of production a month could be managed by 
adding and then laying off some 100 semiskilled men, the saving in inventory 
cost— equivalent to $4,000 per man hired and released— might well justify the 
change. On the other hand, if the change in operating levels involved adding 
and laying off some 1,000 to 1,500 employees of various skills, the inventory 
saving might fall short of offsetting the hiring, training, and layoff costs, not 
to speak of its effect on community relations. Under these circumstances, the 
change might not be worthwhile. 

This alternative production plan, of course, calls for substantially increased 
plant capacity— nearly 60% more— for the same average throughput. If the 
capacity were not available and had to be added, or if it would be gained at 
the cost of overtime or second-shift premiums, or additional equipment in- 
stallations, the simple cost calculation just outlined would have to be extended 
to include these extra costs and investments (not a difficult task if the proce- 
dures are well laid out). 

By making similar trial calculations under other operating patterns, one can 
quickly get a picture of the influence of operating pattern on cost, and can 
arrive at a pattern which comes close to giving the minimum over-all cost. 
This plan then represents the basis for procurement, employment, and inven- 
tory control during the coming months until new forecasts call for an adjust- 

The operating plan summarized in exhibit vi is essentially a minimum-cost 
plan, under the conditions that: (a) inventory costs are $45 per unit; (b) the 
cost of hiring and training an employee is $300 (typical of many industries); 
(c) a change of 750 units in the monthly rate of output requires employment 
or release of 100 men. The cost of seasonal inventory equals 2,150 units 
(average seasonal anticipation stocks) X $45 per unit, or about $97,000. The 

Guides to Inventory Policy 111 


plan calls for varying the production rate from a low of 5,000 units per month 
to a maximum of 11,000 units— a change of 6,000 units; this requires hiring and 
training 800 new employees at a cost of $240,000. (If the hiring and subsequent 
layoff of 800 employees is considered an undesirable employment variation, 
the solution must be sought within whatever are set as the feasible or tolerable 
levels. ) 

Thus, under the plan in exhibit vi the total of seasonal anticipation inventory 
stocks and hiring and training costs is $337,000. This represents a net saving 
of nearly $90,000 per year compared either with the uniform production plan 
or the alternative plan in exhibit iv. (With the hiring and training costs taken 
into account according to the conditions assumed for exhibit vi, the alternative 
plan with its extreme employment variation comes out about the same as the 
uniform plan.) 

Advanced Techniques. Sometimes the problem of planning production 
to meet seasonal demand is too complicated for simple graphical tech- 
niques, and more specialized techniques are needed. One of these is linear 

exhibit vi. minimum over-all cost plan 
(In units) 










safety reserve) 


























1 1,000 





1 1 ,000 







1 2,000 

1 1 ,000 











monthly inventory 


Average safety reserve 


Average seasonal anticipation stock 


programing. The problem just described might have been attacked by 
linear programing methods in order to cut through the repeated trials to 
a good solution, but this approach was not necessary because trial and 
error did not involve a prohibitive amount of time and effort. Linear 
programing has been found useful in circumstances where the problem is 
complicated, for instance, by one or more of these conditions: 

Several product lines using the same facilities or staff. 
Possibilities of planned use of overtime to meet peak needs. 


Inventory and the Administrator 

Need for considering extra-shift premiums. 

Several stages in manufacturing, with seasonal storage possibilities between. 

A number of alternate plants, with different cost and employment situations, 
to meet demand. 

Joint planning of plant operations and of the assignment of branch ware- 
houses to the plant. 

When the seasonal planning problem is attacked as a linear programing 
problem, the objective is to minimize the total of costs incurred in carry- 
ing inventories forward in slack periods to meet future sales peaks, chang- 
ing the production level to meet sales requirements, or resorting to over- 
time. The objective has to be reached within the limitations imposed by: 
(a) capacity restrictions on the amount which can be produced at normal 
or overtime rates in any month; (b) the requirement that inventories in 
each line or product be planned large enough to meet sales requirements; 
and, possibly, (c) the amount of variation that can be tolerated in the 
planned production rate. 

Illustrations of production planning problems formulated in linear pro- 
graming terms can be found in technical literature on the subject. 


Once the production plan has been made, it and the sales forecast 
dictate a sequence of planned inventory balances. However, as sales 
experience accumulates, actual stocks will fall below or exceed the 
planned balances. The minimum inventory balance or safety stock which 
has been (or should have been) set up will absorb the immediate effects 
of departures of actual sales from forecast, but it will be necessary to keep 
adjusting production plans periodically to bring inventories into line. The 
size of the needed safety stock, it should be emphasized, depends on the 
way production adjustments are made. 

The task is comparable to that of adjusting production in the face of 
demand fluctuations, described in the preceding article in this series. There 
it was pointed out that methods used generally take this form: adjusted 
production = original production plan (or forecast sales level) ± some 
fraction or part of the deficiency or excess of inventory compared with 
"normal" or "par." The idea is to keep adjusting production to bring 
inventory back into balance in the face of fluctuations in demand. If the 
fraction is large (close to one), production is made very responsive to 
sales fluctuations, and the inventory needed is smaller. If the fraction is 
small, the inventory acts to absorb sales fluctuations, and must be larger; 
production changes from original plan are smaller. 

Production plans to meet seasonal sales have to be kept in adjustment 
in much the same way, and basically similar control systems can be used. 
In this case, the original plan is the production plan (e.g., exhibit vi) 

Guides to Inventory Policy III 265 

worked out to meet seasonal sales. The "normal" inventory is not a fixed 
level, as in the other case, but varies from month to month; it is the 
planned inventory of exhibit vi, including the safety stock. The steps to 
take in planning production are these four: 

1. From a study of forecast errors or possible differences between sales 
and forecast, and of costs of holding inventory and changing production, 
choose the desired fraction or rate of adjustment in production and the 
corresponding safety stock, using methods of the type described in the pre- 
ceding article. 

The choice of safety stock does not involve production levels, just the costs 
of changing production and holding stocks, along with anticipated forecast 

2. Add the safety stock so chosen to the cumulative sales forecast month 
by month to get the accumulated production required. 

3. Plan production period by period to meet requirements, as described 

4. Periodically adjust the planned production by the specified fraction 
(chosen in Step 1) of the departure of actual inventories from the plan for 
the period. 

For those interested in the actual working through of a problem, 
exhibit vii shows the mathematical expression of the production control 


The relative importance of anticipation stocks and of lot size stocks 
and fluctuation stocks (described in the previous articles) will differ from 
case to case. A study of sales and production characteristics is basic in 


The rule can be written formally as: 

P t = P*± *(Ju-7u);o< k<i 

Pi — adjusted production plan for period i 

P t = original production plan for period i 
Ii_ t = planned closing inventory for period i — 1 

/i_ a = actual closing inventory for period i — 1 
k = fraction of inventory departure adjusted for in production 
# # 

P t and U are chosen by the methods described earlier to minimize total in- 
ventory and operating costs and meet the production requirements: 

Ri = Fi + Si 

266 Inventory and the Administrator 


Ri = production required up through period i 

Fi — accumulated forecast of sales through the period i 

Si = safety stock needed for period i 

The safety stock, Si for each period is in general proportional to expected 
forecast errors and related to the value of k that is chosen. Thus, if forecast 
errors from period to period are independent, 

V ik - k 2 

<ri being the standard deviation of forecast errors for the period; and A a 
parameter which depends on the percentage of customer orders which man- 
agement desires to fill directly from stock (typical values range from 1.3 to 2.5, 
corresponding to 90% and 99% protection). 

finding out what inventory functions are important, and what the sig- 
nificant costs and policies related to these functions happen to be. 


Sales characteristics which strongly influence the production and in- 
ventory control system (and the relative importance of the different in- 
ventory functions) include: 

1. The unit of sales— Are sales made in dozens, tons, or carloads? Planning 
must be done in terms of this characteristic unit. It is obviously not enough, 
for example, to have several tons on hand if the usual unit required is a carload. 

2. The size and frequency of orders— Are there a few large orders each day 
or week, or a steady stream of small orders? This is related to the question of 
unit of sales, but the same total volume sold in a large number of small orders 
can characteristically be supported by substantially less inventory than if sold 
in a few large orders, unless special measures are taken to reduce the un- 
certainty about the time when individual large orders will be placed. 

3. Uniformity or predictability of sales— Do sales show predictable seasonal 
fluctuations? Or do they show large short-term fluctuations, uncontrollable or 
self-imposed (as by special sales campaigns)? Handling large, unpredictable 
fluctuations requires flexibility and additional capacity in inventory production 
as well as carefully designed rules for adjusting or controlling inventory bal- 
ances. But where fluctuations are predictable, advance planning techniques can 
be used. 

4. Service requirements or allowable delay in filling orders— Where allow- 
able delays are small, inventories and production capacity must be correspond- 
ingly greater; care is required to be sure the control system is really responsive 
to needs. 

5. The distribution pattern— Do shipments go direct from factory to cus- 
tomer, through field warehouses, through jobbers, retailers, or consignment? 
The more stages there are, characteristically, the more inventory is required. 
Field inventories in fact serve basically to improve service to jobbers or re- 
tailers and thereby to remove from them some of the burden of keeping 

Guides to Inventory Policy HI 267 

Where the product moves through several stages of handling from factory 
to ultimate consumer, prompt reports or estimates of movement, as close to 
the consumer level as possible, are important in minimizing the amount of 
uncontrollable fluctuation in demand which the factory has to contend with. 
Often the reordering habits of retailers and jobbers can seriously exaggerate 
the basic uncertainty in consumer demand for a product, and thereby com- 
pound the inventory and production control problems of the plant. 

6. The accuracy, frequency, and detail of sales forecasts — Fluctuation stocks 
exist basically because forecasts are not exact. Thus the inventory problems of 
a business are directly related to its inability to forecast sales with precision. 
This does not mean that lack of precise sales forecasts is an excuse for sloppy 
control. Sometimes it is more economical to accept the forecasting uncertain- 
ties and stick to the plan, whether it means overproduction or underproduc- 
tion, than to pay the price in inventories or production fluctuations. But the 
responsibility of forecast errors for inventory needs should be clearly recog- 
nized, and the control system should be adapted to the type of forecasts that 
are possible. 


The production characteristics which influence the scheme of produc- 
tion and inventory control are: 

1. The form of production organization— Job-shop type organization is an 
expensive way of getting flexibility; a company using it should be sure it really 
needs that degree of flexibility. The inventory and production control scheme 
can be considerably simpler under a product-line organization than in a job 

2. The number of manufacturing stages— Where a number of stages in manu- 
facturing exist, the inventory control scheme must be set up to take advantage 
of differences in cost and obsolescence risk which are likely to exist. 

3. The degree of specialization of the product at specific stages— Is each end 
product distinct from the raw material stage on, or are the different products 
more or less the same up to the final processing, assembly, and packaging? 
Where the latter is true, economies are often possible in keeping the right 
balance of stocks in the semifinished state and by simplifying the control and 
scheduling of preliminary stages where the types of product are not diverse. 

4. Physically required processing times at each stage— Processing times affect 
the length of delay, after issuing a replenishment order or adjusting a produc- 
tion rate, before the action becomes effective. The length of this delay, in turn, 
directly influences the size of the inventory needed. 

5. Capacity of production and warehousing stages— Capacity obviously af- 
fects the size and frequency of reorder. 

6. Production flexibility— How rapidly can management vary production 
rates, shift personnel among product lines or departments, and change equip- 
ment from one product to another? Management of inventories and production 
control are basically a question of striking a balance among production flexi- 
bility and capacity, inventory levels, and customer service needs. No company 
is free to pick all three at will. A realistic inventory control system must be 
set up to recognize the limitations in flexibility which exist. 

268 Inventory and the Administrator 

7. Kind of processing— Are batches of materials of a certain size needed in 
production? If so, the quantities and combinations must obviously be taken 
into account in scheduling for production. 

8. Quality requirements, shelf-life limits, or obsolescence risks— These set 
important upper limits on the extent inventories can be used to buy flexibility 
and free production operations from fluctuations in demand. 

These sales and production characteristics cannot be readily distin- 
guished as having one type of effect or another on the production plan- 
ning and inventory control scheme. Nor is it true that one type of char- 
acteristic dictates one approach while another kind of product always 
requires something else. However, the job of setting up a sound produc- 
tion and inventory control system is not just a job of setting up the right 
clerical routines and staff organization; it is a research job to find out 
how the product sales and production characteristics can be exploited to 
get an economical balance between production flexibility, inventory in- 
vestment, and customer service. 


The need for estimates of future sales to control inventories is clearest 
in the case of anticipation stocks, but it exists in the case of the other 
functions as well. Whether forecasts are needed or possible is not the 
question; they are made formally or informally every time a decision is 
made whether to build or replenish an inventory. The question is whether 
the necessary forecasts are being made as well as they might be if formally 
recognized and if available statistical and market research techniques were 
used. Without going into the methods of forecasting, which is a consider- 
able subject of its own, the following points are significant here: 

Economical inventory plans depend on realistic estimates of need— not just 
sales goals or quotas. Even so, there are bound to be forecasting errors— and 
the bigger the possible errors, the bigger the inventories must be to guard 
against them. 

A single forecast figure, without specifying the estimated error or limits of 
error, is not enough. Sometimes the need may be met by a maximum sale 
forecast indicating the upper limit of demand which the production or dis- 
tribution organization will be required to service. 

To estimate the limits of error requires a comparison of past forecasts and 
sales— and often this is hard to do, either because the earlier forecasts were 
made informally, or the records were discarded and hopefully forgotten. 

In any event, forecasting errors bear so importantly on inventory economy 
that to keep the control system up to date requires systematic review of past 
errors and effort to improve the forecasting method. 

Guides to Inventory Policy III 269 


The task of translating inventory policy into practice, of reacting to 
demand as it materializes and utilizing the inventory balances and planned 
production capacity, is a function of production scheduling. Considerable 
effort has gone into the development of techniques— board displays, filing 
systems, card systems, and so on— to facilitate scheduling and control of 
progress on orders scheduled. These techniques can be extremely useful 
if they are adapted to the nature of the product and manufacturing facili- 
ties and if they are used in a framework of self-consistent inventory 
balances and production operating levels. The essence of the control 
problem is setting this framework in the light of management policy, not 
making the actual schedules. 

Conventional scheduling methods are often worked out to cope with 
the complexities of job-shop production, where each order is unique and 
no set sequence of operations exists. 

Scheduling operations in this way through a large number of stages or 
departments is difficult. Fortunately, however, the need for so doing is 
not nearly so common as one might gather. Many businessmen, in dis- 
cussing inventory and production control, give the impression that their 
organization with its large product line— whether several hundred or 
several tens of thousands— is saddled with job-shop operations from top 
to bottom. They look with longing toward the lower operating costs and 
simpler management problems of assembly-line operations. They fre- 
quently fail to recognize that almost all products and product lines are 
capable of being manufactured under a wide range of organizational 
forms intermediate to the extremes of either pure job-shop or assembly- 
line operation. 

Taking advantage of this latitude has been a source of considerable 
operating economy in some businesses— and could be in many more. 


A comprehensive inventory control system should be closely co-ordi- 
nated with other planning and control activities, such as sales forecasting, 
cash planning, and capital budgeting, since it affects all of these activities 
in many ways. The specific steps and timing will vary from one company 
to another, depending on product and process requirements, but the es- 
sentials of an inventory control system can be grouped into three broad 
classes: long-range inventory planning, short-range planning, and schedul- 


Inventory and the Administrator 



















_ 1 



< p= o 



O . 


Q£ I— 


Guides to Inventory Policy HI 271 

exhibit viii shows the three basic planning functions in boxes, with the 
arrows indicating the flow of information to and from the analysis. More 

1. The long-range plan makes use of: (a) sales forecasts, with error or 
range estimates, and (b) preliminary policy decisions on capital allocation 
and value and on the amount of risk to be assumed. The purpose is to show the 
implications of policy choices so they can be refined and sharpened, and then 
to provide a basis for long-range operating decisions concerning construction, 
purchase, and sale of facilities, adjustment of sales and promotion programs, 
and so on. The analysis results may also lead to further forecasting effort by 
showing the production and capital costs resulting from poor forecasts. 

2. At the intermediate stage, the short-range plan uses as its "raw materials" 
or inputs: (a) the results of policy decisions, (b) short-term demand fore- 
casts, (c) existing facilities and manpower, and (d) inventories. The outputs 
are bases for short-range operating decisions— the general production plan to 
follow, adjustments in the employment rate, corrections in inventory bal- 

3. Finally, within this framework scheduling can react to demand as it 
actually materializes. 




Edward P. Burnham* 

Mr. Burnham gives a step by step program for assuring a imiform and 
realistic gathering of data for research program and cost analysis. 


Ralph H. Manley has defined the purpose of research in terms of a 
responsibility of management "to provide the technical leadership neces- 
sary in order for the company to earn a satisfactory return on its invested 
capital, both this year and especially down through the years to come." 

Restated as objectives, this responsibility becomes: 

1. To maintain the company's prestige and profits by keeping existing 
products competitive in quality and price. 

2. To improve the company's competitive position and increase profits 
by developing new products that replace or supplement existing products, 
and by improving present products to a point where they have greater ac- 
ceptability in the market. 

* From The Management Review, XL VII, 8 (1958), 20-24, 80-83. Reprinted by per- 
mission of the American Management Association, Inc. and The Management Review. 


Controlling the Costs of Research 273 

3. To explore possibilities for expansion into related or unrelated fields 
that offer opportunity for substantial profits. 

If we accept these as the objectives of research, then we have a charter 
for the establishment of budgets and cost reports, which in turn will pro- 
vide data for control of research by management. 


Costs must be measured on a consistent and well-defined basis before 
they can provide meaningful data for management decisions. This re- 
quires a clear understanding of what is to be included in-and excluded 
from— research costs. 

Fundamental research should be included, since it covers work leading 
to new technology, even though it has no particular connection with 
present products. In addition, projects leading to entirely new products 
or processes should be included when they are within the broad scope of 
the company's present field of activity. And finally, projects should be 
included when through application of new technology, they lead to 
improvement of present products and processes, or when they are de- 
signed to protect current investment and market position. 

At the same time, some items seem to find their way into the research 
budget even though they do not belong there. Examples of this kind of 
cost include technical advice or assistance to help the production depart- 
ment out of difficulty or to carry on its normal operations; trouble- 
shooting to correct production errors that have reduced normal standards 
of materials; cost-reduction activities for the production department; 
technical advice or service to the marketing department on specific cus- 
tomer problems; assistance to the marketing department in the extension 
of going products into new application areas; and preparation of samples, 
at marketing department request, for specific customer requirements. 

Since the research department must account for and justify its expendi- 
tures in terms of results, it should only be charged with the costs of 
research activities. All work of a service nature should be charged to the 
department for which it is performed, and the accounting system should 
include a method for separating these costs from research costs. The ex- 
tent of the research department's service effort should be measured by the 
availability of research personnel to perform these services and still fulfill 
the research mission. 

We must also decide when research responsibility-and therefore re- 
search cost-ends. Depending on established policies, any one of the fol- 
lowing guides can be used: 

When production accepts the process. 
When commercial sale begins. 

274 Research and Development 

When the product is made in interim facilities. 

When a pilot plant produces small quantities for sale. 

When the product is turned over to the engineering department for com- 
mercial design. 

When production drawings, a working model, and standard manufactur- 
ing practices are complete. 

When the product has been manufactured long enough to show that it can 
be produced in quantity. 

When the product is transferred to a manufacturing company or product 

Nearly all research departments provide technical service for other 
departments, and some companies assign full responsibility for solving 
all problems of a technical nature to the research department. However, 
the majority of companies provide technical personnel in other depart- 
ments—production, sales, engineering, etc.— to allow research personnel 
to concentrate their efforts on new products and processes. 

When charges are made to other departments, they are accumulated 
under appropriate project numbers, and the total cost is transferred at 
periodic intervals from research accounts to the department requesting 
the service. The practice of making interdepartmental charges reflects 
their costs accurately and does not burden research with what properly 
should be manufacturing or sales expense. Furthermore, this practice re- 
quires that other departments justify their requests and discourages undue 
demands on the research department. 

Interdepartmental charges are not made when it is established that the 
work is primarily research, when other departments might otherwise be 
duplicating existing projects, or when the company decides to minimize 
the expense of accounting for internal transactions. 


Some large research departments have separated technical activities 
from administrative functions such as budgeting, accounting, and servic- 
ing. This permits research management and technical personnel to con- 
centrate on research projects. When this approach is taken, the business 
administrator is usually a member of the research staff and reports directly 
to the director of research. He prepares budgets, assists in maintaining and 
improving the plan of control under the budgetary program, and guides 
or co-ordinates accounting reports. 

When research departments are small, budgeting and accounting are 
normally performed by the accounting department. Regardless of organ- 
ization, however, the director of research is responsible for financial 
plans embodied in the research budget. 

The plan for classifying costs should meet management requirements at 
all levels, providing the necessary data from current accounting records. 

Controlling the Costs of Research 275 

The same classifications will, of course be used for budgeting as well as 
for recording current costs. 

What does management want to know? If we consider managements 
requirements as questions, then our classifications system can be designed 
to answer them. 

What was the total spent for research? This total can be obtained by 
having a single account to which all research expenses are charged. Since 
both the research director and top management are concerned with this 
type of data, it might be desirable to separate the expenses of different 
laboratories. The basic control account can remain the same, with the 
use of additional code numbers to identify each laboratory. 

What was the cost of each individual project? When it has been de- 
cided to activate a particular research project, a project number is as- 
signed and all expenses are charged to it. The research director and group 
leader can then determine the total expense for any project for a given 
period, thus measuring the research effort in terms of cost to the various 

What was the cost of the items making up the total expense? Normally, 
we can use the same expense elements, codes, and accounts that are as- 
signed to other departments of the company, though some expenses 
peculiar to research will require special accounts. 

Expense elements such as salaries, wages, supplies, depreciation, in- 
surance, taxes, travel, and communications are common to all departments. 
However, items such as library, professional services, speakers, member- 
ships, subscriptions, consultants, and outside research are often peculiar to 

The research director and group leader can assign expenses to the 
proper code and determine the total expense by element for all projects 
and for each individual project. 

Who spent the money? The answer to this question is an easy one, since 
the project was assigned to an individual group or area within the depart- 
ment when it was approved. By establishing a code to designate the 
various laboratory subdivisions to which projects are assigned, the research 
director and group leader can identify the spender. 

How much was spent by type of research? Projects may call for work 
in application research, pilot plant research, outside research, synthesis, 
new methods and techniques, or any of a number of other areas. Records 
of these expenses are important to the research director and top manage- 
ment to permit control and direction and to determine the total effort or 
dollars expended in each category. 

276 Research and Development 

How 77221 ch was spe7it by class of research? When a project is approved, 
it must be classified as to whether it is designed to support present invest- 
ment and sales, to lead to new products and processes for the future, or 
to advance our fundamental knowledge. Coded expenses provide data to 
show management and the research director how effort or dollars are 
being apportioned between the present and the future. 

How much was spent for individual products? In companies with 
several product areas, management and the research director may wish to 
know the expense associated with each product. Each project is classified 
according to the product area. Expenses of projects applicable to more 
than one area or those of a general nature that benefit all areas can be 
charged to a special category and later prorated against established areas. 


By considering the questions which management wants answered we 
have developed a plan for classification and coding of research expenses. 
A typical project classified under such a coding system might be listed as 


Question Heading Assigned Number 

What was the total spent? Department 412 

What was the cost of the project? Project Number 2983 

What type of expense was it? Expense Element 218 

Who spent the money? Group or Individual 34 

What was the expense by type Type of Research 1 2 

of research? 

What was the expense bv class Class of Research 02 

of research? 

What was the expense for a product? Product Area 07 

The code would then appear as: 412 2983 218 34 12 02 07. Several code 
numbers— group or individual, class of research, and product area— will 
not change during the life of the project. Those that do must be included 
by the laboratory employee in his time report. 

With this information, the accounting department can prepare reports 
that furnish short-term records for research management and top manage- 
ment, serve as a basis for budgets, and furnish a long-term record of cost 
by projects, type and class of research, and product area. These tech- 
niques can also be applied to other classifications that may be of impor- 
tance in research cost control. Of course, any of them may be omitted 
when they do not apply. 

This system is an effective method of handling direct expenses. In 
addition, overhead expense must also be charged to these projects. One 
method of accomplishing this is to prorate the overhead expense in rela- 
tion to direct salaries and wages. 

Controlling the Costs of Research 277 


Management controls research activities by limiting the budget with 
appropriated funds. Many companies use a fixed percentage of their sales 
to determine the research budget, and, when the sales forecast fluctuates 
downward, the research department is forced to reduce its planned pro- 
gram, usually at the expense of long-range projects. In some cases, this 
could prove detrimental to the future growth of the company. 

A budget based on plant investment is more stable and presently is 
being used by some companies. This plan must remain flexible, since un- 
usual expenditures and new development projects will require special 


Since technical personnel represent the largest single item in the re- 
search budget, they can be used as a cost base. A study of the year's 
program will reflect the proportion to be devoted to maintenance research, 
new-products research, and fundamental research, while detailed analysis 
of the individual projects will determine the manpower requirements. 
Finally, the cost per technical man can be determined from the various 
expense elements such as salaries, wages, and supplies. Preparation of 
other budget schedules can then follow rather easily, since the number of 
technical people assigned to various projects and classification areas is 
known. This type of computation may not always be completely ac- 
curate, since some groups may require fewer supplies or repairs and 
unpredictable factors can always enter into any of the projects. 

Since other departments are normally charged for work of a service 
nature, a section must be established in the budget for credits. This can 
consist of items for each type of service-such as service to the produc- 
tion, sales, and engineering departments-or all may be included in one 

In effect, then, a gross and a net budget are developed. It is the net 
budget that determines the scope of the research effort, since it is within 
its limits that actual research projects are accomplished. 

Once expenses are classified, a budget schedule can be developed to 
match them. Periodic reports from the accounting department will then 
show actual expense versus budgeted expenses for each item on the 
schedule. This consolidated report provides the basic information needed 
for control of the research program. 

With these data, management can readily determine whether the re- 
search department is following the program as budgeted, and can modify 
or shift activities as necessary. 


Planning is the key to control. Planning on the company level deter- 
mines the amount of money that can be allocated to research, and thereby 

2 y8 Research and Development 

sets the over-all budget. Planning by research management distributes 
these budget dollars to fit the particular program. And at other super- 
visory levels, planning is the key to efficient use of this money in pursuing 
work on specific projects. 

Budget schedules are developed to reflect this planning, and cost ex- 
pense reports are developed to measure performance on an expense basis 
and to provide means of controlling these expenditures. 

With proper budgeting and cost-reporting procedures, management 
has the tools to analyze, control, and schedule research activities to the 
best future interests of the company. 



Adolph G. Lurie* 

Mr. Adolph G. Lurie lists and discusses the factors of importance when 
using the budgetary process to control research costs. 

The budgeting and cost control principles used in effective management 
of the manufacture and sale of products can also be used by management 
in controlling the expenditures for research and development. 

The establishment of a research and development budget in total is a 
relatively simple matter. However, the preparation of such a budget for 
effective cost control is a more comprehensive procedure. The methods 
which may be followed for both steps will be outlined here. 

Budgets for research and development can be established by several 
different methods depending upon the viewpoint of top management. 
Suggested methods for determining the amount of dollars to be expended 
during a given period, usually a year, can be based upon any one of the 
following considerations, or any combination of them that management 
may desire to recognize: 

1. The total amount of the budget may be based upon the sales for the past 
period or it may be a fixed percentage of the estimated sales for the ensuing 

2. It may be desirable to budget a percentage of net profits before taxes. 

3. Management may decide to base the budgeted expenditure upon the 

* From National Association of Cost Accountants Bulletin, XXXIV, 7 (1953), 
894-901. Reprinted by permission of the National Association of Accountants. 

Controlling Costs with a Budget 279 

amount that had been spent previously, modified either upwards or down- 
wards by changes in volume of sales, changes in profits, or similar considera- 

4. The research budget may be dependent upon the operating budget and 
the amount determined from the forecast of sales or upon budgeted profits 
before research and development. 

5. A general review and study of economic conditions, future prospects, 
competition, etc. may influence the establishment of the budget. 

6. The least scientific method of approaching this problem is to fix the 
amount by arbitrary determination. 

Thus the total amount of the budget for research and development is 
established in one or more of these ways and the first major premise for 
the budgetary and cost control of a research and development department 
is provided. It is just a starting point for this purpose. Merely establishing 
a budget does not give the directors of the research department any guide 
for programing their efforts. Such guidance is invaluable for those who 
must make decisions, direct the orderly progress of the development pro- 
gram, and intelligently plan and control the work of the department. 

The preparation of the budget is a twofold operation, consisting of 
setting up a budget classified by types of expenditures, to enable the 
technical director and supervisors to provide for facilities and staff in 
accordance with the amount of money available (exhibit i) and also a 
budget for projects contemplated during the period (exhibit 2). The 
order of preparation of these two budgets depends upon management's 
approach to the problem but both budgets are for the same dollars. These 
budgets should be a joint project of the research department and the 
budget or cost accounting department, to insure compliance with tech- 
niques established for the accounting of the expenditures. The first of 
the two arrangements of the budget is relatively easy to prepare and can 
be fairly definitely fixed. The second, which breaks the total up by 
projects, should provide for changes throughout the period to allow the 
technical director and his supervisory staff flexibility in their operations. 


The budget of expense classifications may be fairly simple hr form but, 
if a comprehensive budget is desired, can also be quite involved. The 
degree of complexity depends largely upon the size of the department, 
the amount of money involved and the amount of information and 
control desired by those responsible for managing the research depart- 
ment. A basic budget may consist of: 

1. Pay rolls— salaries and wages, including related costs, such as social secu- 
rity taxes, compensation insurance, group insurance, pensions, etc. 

2 8o Research and Development 

2. Supplies and materials, such as expendable equipment and operating sup- 

3. Other direct operating costs. 

Such a grouping, comprising only a few figures, may be adequate for 
a satisfactory budget and cost control of the expenditures of a small- or 
medium-sized department. A description of the elements entering into 
each of the above classifications is shown in more detail in the comprehen- 
sive budget outline below: 

1. The subdivisions under payrolls, may be the following: 

a. Salaries of professionally trained personnel. 

b. Salaries and wages of nonprofessional employees, such as laboratory 
technicians, draftsmen, etc. 

c. Salaries of service employees, usually stenographers, and clerical work- 
ers of the department. 

d. Plant labor, consisting of hourly workers borrowed from operating de- 
partments for specific work as required. 

2. Supplies and materials consist of two major items, namely: 

a. Expendable equipment purchased for specific projects or for general 
use in the research department, which does not become a part of the 
basic equipment. 

b. Supplies and materials. 

3. Other direct costs can include, but need not be limited to the following: 

a. Books, periodicals, dues, subscriptions, and similar items for the de- 
partment staff or library. 

b. Travel expense. 

c. Fees for outside technical, engineering, and consulting services. 

d. Taxes, depreciation, and insurance on building, permanent fixtures, 
furniture, and equipment. 

e. Cost of service facilities from the plant, such as light, heat, power, 
steam, etc. 

f. Miscellaneous. 

The largest item in a research department budget is usually the salaries 
and wages of those employed in the department. It is relatively simple to 
develop the dollar amounts by totaling the salaries to be paid to the indi- 
viduals engaged in research work, including payroll taxes, insurance and 
the cost of other benefits. An estimate can be made of the plant labor re- 
quired in the operation of the research and development department, 
based upon a study of previous experience. 

The budget for supplies, materials, and expendable equipment can also 
be based upon past experience, adjusting for changes in present require- 
ments and the relative cost of these materials. Under present conditions, 
even though the quantity required may be the same as for the previous 
year, the cost would rise due to increased prices. 

Other direct costs can be readily determined from the factors involved. 
Dues and subscriptions are relatively fixed. The books to be purchased for 
the library can be based upon past experience. Travel expense depends 

Controlling Costs with a Budget 



Expense Budget 
Year ended December 31, 1953 

1. Payroll 

a. Technical employees $150,000 

b. Non-technical employees 30,000 

c. Service employees 20,000 

d. Plant labor 10,000 

Subtotal 210,000 

2. Supplies and materials 

a. Expendable equipment 25,000 

b. Operating supplies 15,000 

Subtotal 40,000 

3. Other direct costs 

a. Books, dues, subscriptions, etc. 5,000 

b. Travel expenses 15,000 

c. Technical, engineering, and consulting fees 15,000 

d. Taxes, insurance, depreciation 5,000 

e. Light, heat, power, etc. 5,000 

f. Miscellaneous 5,000 

Subtotal 50,000 

TOTAL BUDGET $300,000 

upon several factors, such as the number and location of principal meetings 
of technical societies, the locations of the plants of the organization, the 
area in which the customers are located, and the general company policy 
with regards to travel expense. Past experience can be a good guide in 
determining what the expenses might be in the future. 

Technical, engineering, and consulting fees can be ascertained from a 
review of the projects contemplated during the period. The capabilities of 
the research personnel and the contemplated changes therein must be 
considered in arriving at the amount of money that would be spent for 
outside assistance. 

Taxes, insurance, depreciation, light, heat and power and other similar 
expenses are usually fixed by the distribution of the total cost to the com- 
pany and the amount to be allocated to the research and development de- 
partment would, therefore, be obtained from the budget of the operating 


Research and Development 


Consideration may also be given to setting a policy with respect to 
credits from the operation of the research and development department 
resulting from: 

i. The sale of finished materials produced by the research and development 

2. The sale of unusable expendable equipment or scrap resulting therefrom. 

3. The transfer of expendable equipment or other materials to operating de- 

4. Charges to customers and others for technical services. 

5. Any other miscellaneous credits, depending upon the accounting policy 
of the organization. 

In a situation in which sales may be sizable in amount and also regular, it 
would be advisable to establish a section in the budget for credits. Thus, 
management can control the situation and see that all possible credits are 
given to the development department. However, where sales resulting 
from research and development work are limited in relation to the entire 
research program, it would be far simpler not to include such items in the 
budget but to allow the division to dispose of materials as they see fit, 
providing an incentive to sell unusable equipment and materials, thereby 
increasing funds for completion of projects. 

exhibit 2 


Project Budget 
Year ended December 31, 1953 

1 . Present products 

Current Prior Author- 

Budget Expenditures ized 

a. Projects in progress 

( 1 ) Products x improvements 

(2) Product y usage 

( 3 ) Product z quality 

$ 5,000 





$ 20,000 




b. New projects 

(1) Product x new process 

(2) Product y quality control 

(3) Product z new use 






Total present products 




Controlling Costs -with a Budget 


2. New product research 

Total Sales Department service 

5. Service departments 

a. Library 

b. Drafting room (general) 

c. Stenographic and clerical 

6. Balance for unauthorized projects 










a. Projects in progress 

( 1 ) Product xx 

(2) Product yy 








b. New projects 

(1) Product P 

(2) Product Q 

(3) Product R 







Total new products 

3. Pure research 

a. Projects in progress 

(1) ItemS 

(2) Item T 






1 2,000 



b. New projects 

(1) ItemU 

(2) ItemV 






Total pure research 

4. Sales department service 

a. Projects in progress 

( 1 ) Product x 

(2) Customer z 











b. New projects 

( 1 ) Product y 

(2) Product q 













We have discussed the establishment of a budget by classes of expendi- 
tures. However, the important phase of research budgeting for cost con- 
trol is the preparation of a budget for each project to be undertaken 

284 Research and Development 

during the period. This likewise can be relatively simple or quite elaborate, 
depending upon the degree of management cost control desired. The 
simplest procedure would be merely to list the projects which will be 
undertaken, showing the amount of the budget allocated to each. Pro- 
visions should be made for additions to the list of projects during the 
period as new fields of research are entered. The total of the amounts 
authorized for the several projects must agree with the total of the amount 
budgeted by classes of expenditures. 

A more detailed approach to budgeting research projects might be one 
in which projects are grouped under several headings. Major classifications 
could be as follows: 

1. Improvement in the manufacture, quality, and usage of present products. 

2. Research and development of new products. 

3. Projects requested by customers or the sales department. 

4. Pure or "blue sky" research, having no commercial value. 

5. Service departments, research library, clerical, stenographic, drafting 
room, etc. This should provide only for that portion of research and develop- 
ment expense not directly allocable to projects. 

6. Balance available for the projects to be authorized at a later date. 

Under the first four of the above classifications, there could be a 
specific budget authorized for each project, in groups as follows: 

1. Completion of projects authorized in prior periods. 

2. Projects to be started in current period. 

exhibit 2 is a sample form which could be followed in setting up this 
portion of the budget. As previously stated, the total of all the items 
entered on it must agree with the total expense classification budget. 

A logical starting point is a determination of the amount necessary to 
complete projects already in progress. Another portion of the budget 
easily determinable is the cost of operating the research service depart- 
ments. These two factors, deducted from the total budget, result in the 
balance available for new projects. The research directors and supervisors 
can then determine the portion of this remainder to be allocated towards 
research relating to products presently being produced, new products, or 
for pure research. The manner in which the totals for each of these major 
classifications is determined, depends upon the nature of the business and 
the direction in which research and development effort is to be expended, 
based upon the policy of the organization. 

Following the determination of the amount for each major classification, 
the next step is an apportionment to the individual projects being con- 
sidered for active investigation. In this apportionment, no attempt should 
be made to show the manner in which funds are to be expended, since the 
amount of materials, labor, or supplies required for each project cannot 
be determined readily and must be ascertained as the project progresses. 

Controlling Costs with a Budget 285 

Under the procedure outlined, management authorizes the research and 
development division to spend a specified sum in its over-all operation and 
also specifies the amount which may be spent for the projects to be under- 
taken. Efforts to pinpoint the budget to the extent that a figure for the 
several classifications of expenditure for each project is developed, might 
so limit the research department as to interfere with its smooth operational 

The outlined procedure provides sufficient flexibility so that the direc- 
tors of research may use their own judgment in shifting the efforts within 
their division among the various projects. In this connection, it is advisable 
to revise the budgets periodically by issuing supplementary budgets. Un- 
used balances may be transferred from projects which are completed to 
"balance for unauthorized projects," to new projects, or to projects in 

As shown in exhibit 2, the project budget, it is desirable to indicate the 
amount expended before the current period, it being assumed that such 
amounts plus the budget for the current period, will be the total amount 
authorized to date for each project. The total budget for the individual 
project is an extremely important guide for management, since control of 
the total expenditure should be maintained to decide whether a project 
should be abandoned or continued to its successful conclusion. It should 
be recognized that a long-range program must be established and pro- 
visions should be made so that the full program can be completed without 
capricious change of policy on the part of financial executives. 


Upon the completion of the expense and project budget, approval by 
top management is desirable and usually required. A letter of transmittal 
of the budget, outlining the high spots and indicating the anticipated 
results of the research effort, is quite helpful for obtaining a prompt ap- 
proval. In preparing the budget, a realistic and practical approach should 
have been used, so as to enable management readily to foresee the bene- 
ficial results of the proposed efforts. The budget should not be top-heavy 
on pure research or nonproductive work and, depending upon the nature 
of the business, should have sufficient stress upon the improvement in 
present products, either from the production, quality, or use viewpoint 
and upon the development of new products. This is desirable to obtain 
the proper interest, enthusiasm and approval of top management. 

The budgeting procedure described here is of value only if actual ex- 
penditures are properly recorded in accordance with the budget, and 
comparisons made between the amounts actually expended and those 
authorized. Cost records are advisable, as well as monthly or quarterly 

286 Research and Development 

cost reports in form similar to the budget, so that those responsible for the 
functioning of the research department can control the expenditures and 
determine the amount of funds available for subsequent operations. With 
such budgeting and reporting, the research director can plan for the future, 
and financial executives can assist in the direction of the research and de- 
velopment division towards the ultimate goal of all concerned namely, 
progress for the organization, growth, and additional profits. 


Cyril C. Herrmann and John F. Magee* 

The scientific method— as practiced in physics, biology, and chemistry 
—is the basis of a new approach to business problems. The article describes 
several operations research applications and discusses the factors affecting 
their use in business. 

There is a new concept in management. It is called operations research. 
It has helped companies to solve such diverse business problems as direct- 
ing salesmen to the right accounts at the right time, dividing the advertis- 
ing budget in the most effective way, establishing equitable bonus systems, 
improving inventory and reordering policies, planning minimum cost 
production schedules, and estimating the amount of clerical help needed 
for a new operation. 

Operations research makes possible accomplishments like these and 
many others because (a) it helps to single out the critical issues which 
require executive appraisal and analysis, and (b) it provides factual bases 
to support and guide executive judgment. Thus, it eases the burden of 
effort and time on executives but intensifies the potential of their decision- 
making role. In this sense operations research contributes toward better 

What is this thing called operations research? How does it work? How 
does it differ from other services to management? Where can it be used? 
How should management get it organized and under way? What are its 
limitations and potentials? These are all questions that we shall try to an- 
swer in the following pages. 

* From Harvard Business Review, XXXI, 4 (1953), 100-112. Reprinted by permis- 
sion of the Harvard Business Review. 

Operations Research 2 ° 7 


Operations research apparently means different things to different 
people. To some businessmen and scientists it means only the application 
of statistics and common sense to business problems. Indeed, one vice 
president of a leading company remarked that if his division heads did not 
practice it every day, they would not last long. To others it is just another 
and perhaps more comprehensive term for existing activities like market 
research, quality control, or industrial engineering. Some businessmen 
consider it a new sales or production gimmick; some, a product of aca- 
demic people interfering in the practical world. In truth, operations re- 
search is none of these things, as we shall soon see. 

It should not be surprising that there has been this confusion. Operations 
research is not an explicit, easily identifiable concept that developed to 
meet the specific needs of industry. It was first applied in World War II 
by groups of scientists who were engaged by the government to help 
frame recommendations for the improvement of military activities. After 
the war a few soundly managed companies experimented with it and 
found that it worked successfully in business operations as well; and it has 
since gained a secure foothold in industry. 

Early attempts by operations analysts to describe their activities, based 
on the objective of arriving at a precise and comprehensive definition of 
operations research, tended to be overly generalized, broad, and self- 
conscious, and suffered from emphasis on military applications. Some of 
the confusion surrounding the meaning of the term, operations research, 
has resulted from attempts at identification with special techniques or 
unnecessarily rigid distinctions between operations research and other 
management service activities. 

Now, let us see if we can cut through some of this confusion. 

The first point to grasp is that operations research is what its name 
implies, research on operations. However, it involves a particular view of 
operations and, even more important, a particular kind of research. 

Operations are considered as an entity. The subject matter studied is 
not the equipment used, nor the morale of the participants nor the physical 
properties of the output; it is the combination of these in total, as an eco- 
nomic process. And operations so conceived are subject to analysis by the 
mental processes and the methodologies which we have come to associate 
with the research work of the physicist, the chemist, and the biologist- 
what has come to be called "the scientific method." 


The basic premise underlying the scientific method is a simple and 
abiding faith in the rationality of nature, leading to the belief that phe- 
nomena have a cause. If phenomena do have a cause, it is the scientist's 

288 Research and Development 

contention that by hard work the mechanism or system underlying the 
observed facts can be discovered. Once the mechanism is known, nature's 
secrets are known and can be used to the investigator's own best advan- 

The scientist knows that his analogue to nature will never be entirely 
perfect. But it must be sufficiently accurate to suit the particular purposes 
at hand; and, until it is, he must repeat the processes of observation, in- 
duction, and theory construction— again and again. Note that a satisfactory 
solution must be in quantitative terms in order that it can be predictive— 
the only accepted fundamental test of being physically meaningful. 

The scientific method, in its ideal form, calls for a rather special mental 
attitude, foremost in which is a reverence for facts. Of course all modern 
executives are accustomed to using figures to control their operations. 
But they are primarily concerned with results and only secondarily with 
causes; they interpret their facts in the light of company objectives. This 
is a much different attitude from seeking out the relationships underlying 
the facts. 

Thus, when an executive looks at sales figures, he looks at them primarily 
in terms of the success of his sales campaign and its effect on profits. By 
contrast, when the scientist looks at these same figures, he seeks in them a 
clue to the fundamental behavior pattern of the customers. By the process 
of induction he tentatively formulates a theoretical system or mechanism; 
then by the inverse process of deduction he determines what phenomena 
should take place and checks these against the observed facts. His test is 
simple: Does the assumed mechanism act enough like nature— or, more 
specifically in this case, does it produce quantitative data such as can be 
used for predicting how the customers will in fact behave? For example: 

In a company manufacturing specialty products, examination of account 
records showed that customer behavior could be accurately described as a 
time-dependent Poisson process— a type of phenomenon found widely in 
nature, from problems in biology to nuclear physics. This concept yielded the 
key to establishing measures of the efficiency of the salesmen's work and of the 
effect of the promotion in building sales. On this basis a new method of di- 
recting promotional salesmen to appropriate accounts was constructed— and 
then tested by careful experiments, to see if sales increases resulted at less than 
proportionate increases in cost. (The results in this case were spectacular: an 
over-all sales rise in six figures, and a corresponding gain in net profits.) 


Through the years mathematical and experimental techniques have been 
developed to implement this attitude. The application of the scientific at- 
titude and the associated techniques to the study of operations, whether 
business, governmental, or military, is what is meant by operations re- 

Operations Research 289 

Newton was able to explain the apparently totally unrelated phenomena 
of planetary motion and objects falling on the earth by the simple unifying 
concept of gravity. This represented a tremendous step forward in helping 
men to understand and control the world about them. Again, more re- 
cently, the power of the scientific method was demonstrated by the 
ability of the nuclear physicists to predict the tremendous energy poten- 
tial lying within the atom. 

Here are a few summary examples of the way this same kind of ap- 
proach has been applied to down-to-earth business problems: 

A company with a number of products made at three different locations 
was concerned about the items to be produced at each location and the points 
at which the items would be warehoused. Freight costs constituted a substan- 
tial part of the delivered cost of the material. Operations research showed that 
what appeared to be a complex and involved problem could be broken into a 
series of rather simple components. Adaptations of linear programing methods 
were used to find the warehousing schedule which would minimize freight 
costs. The study is now being extended to determine the best distribution of 
products among manufacturing plants and warehouse locations in order to 
minimize net delivered cost in relation to return on investment. 

A manufacturer of chemical products, with a wide and varied line, sought 
more rational or logical bases than the customary percentage of sales for 
distributing his limited advertising budget among products, some of which 
were growing, some stable, and others declining. An operations research study 
showed that advertising effectiveness was related to three simple characteristics, 
each of which could be estimated from existing sales data with satisfactory 
reliability: (a) the total market potential; (b) the rate of growth of sales; 
(c) the customer loss rate. A mathematical formulation of these three charac- 
teristics provided a rational basis for distributing advertising and promotional 

In a company making a line of light machines, the executive board ques- 
tioned the amount of money spent for missionary salesmen calling on custom- 
ers. Studies yielded explicit mathematical statements of (a) the relation be- 
tween the number of accounts called on and resulting sales volume and (b) the 
relation between sales costs and manufacturing and distribution costs. These 
were combined by the methods of differential calculus to set up simple tables 
for picking the level of promotion in each area which would maximize com- 
pany net profits. The results showed that nearly a 50% increase in promotional 
activity was economically feasible and would yield substantial profits. 

An industrial products manufacturer wanted to set time standards as a basis 
for costs and labor efficiency controls. The operations research group studied 
several complex operations; expressed the effect of the physical character- 
istics of products and equipment and the time required to produce a given 
amount of output in the form of mathematical equations; and then, without 
further extensive time study or special data collection, set up tables of produc- 
tion time standards according to product characteristics, equipment used, and 
worker efficiency, which could be applied to any or all of the production 

A company carrying an inventory of a large number of finished items had 
trouble maintaining sound and balanced stock levels. Despite careful attention 
and continued modification of reorder points in the light of experience, the 

2 go Research and Development 

stock of many individual items turned out to be either too high for sales or 
inadequate to meet demand. The problem was solved by a physical chemist 
who first collected data on the variables, such as size and frequency of order, 
length of production and delivery time, etc.; then set up an assumed system, 
which he tried out against extreme sales situations, continually changing its 
characteristics slightly until it met the necessary conditions— all on paper (a 
technique well known to physical scientists); and thus was able to determine a 
workable system without cost of installation and risk of possible failure. 

These examples should serve to give some idea of how the scientific 
method can be applied. But they represent only a few of the many scien- 
tific techniques available (as we shall see when we examine further cases 
in more detail). Some practitioners even take the rather broad point of 
view that operations research should include the rather indefinite and 
qualitative methods of the social fields. Most professional opinion, how- 
ever, favors the view that operations research is more restricted in mean- 
ing, limited to the quantitative methods and experimentally verifiable 
results of the physical sciences. 


There are four concepts of fundamental importance to the practice of 
operations research: (a) the model, (b) the measure of effectiveness, (c) 
the necessity for decision, and (d) the role of experimentation. 


The most frequently encountered concept in operations research is 
that of the model— the simplified representation of an operation, contain- 
ing only those aspects which are of primary importance to the problem 
under study. It has been of great use in facilitating the investigation of 
operations. To illustrate with some familiar types of "models" from other 

i. In aeronautical engineering the model of an aeroplane is used to investi- 
gate the aerodynamic properties in a wind tunnel. While perfectly adequate 
for this purpose, it would hardly do for practical use. It has no seats; it may not 
even be hollow. It is, however, a satisfactory physical model for studying 
the flight characteristics of the ship. 

2. Another, quite different kind of model, with which we are all familiar, 
is the accounting model. This is essentially a simplified representation on 
paper, in the form of accounts and ledgers, of the flow of goods and services 
through a business enterprise. It provides measures of the rate of flow, the 
values produced, and the performances achieved, and to that extent is use- 
ful (though it is hardly a realistic representation of operations). 

3. Many models are used in physics. Three-dimensional models of com- 
plex molecules are probably most familiar to laymen, but the most power- 
ful models in this field are sets of mathematical equations. 

Operations Research 2 9 ! 

There are several different types of operations research models. Most 
of them are mathematical in form, being a set of equations relating sig- 
nificant variables in the operation to the outcome. An example, illustrating 
how a company used a model to improve time standards on the production 
line, is provided in exhibit i. 

Another type of model frequently used is the punched-card model, 
where components of the operation are represented by individual punched 
cards; masses of these are manipulated on standard punched-card equip- 
ment. For example, in a study of a sales distribution problem, each cus- 
tomer, of thousands served by the company, was represented by a 
punched card containing significant information about his location, type 
of business, frequency of purchase, and average rate of business. The 
punched cards representing the customers could then be subjected to as- 
sumed promotional treatments, with the effects of the promotions punched 
into the cards. The resulting business could be calculated and an evaluation 
made of alternative sales-promotion campaigns. 

Occasionally a model is physical like the ones often used by engineers. 
For example, the use of a hydrokinetic model has been proposed in the 
study of a mass advertising problem. The fluid flowing through the model 
would represent business of various types going to the company or to 
competitors as a result of various forms of the company's own and com- 
petitive promotional efforts (represented in the model by forces acting 
on the fluids). 

Operations research models can also be distinguished as exact or prob- 

i. An exact model is used in operations or processes where chance plays 
a small role, where the effect of a given action will be reasonably closely de- 
termined. Exact models can be used, for example, in long-range production 
scheduling problems in the face of known or committed demand. The exact 
model is sufficiently accurate since it can be assumed that, barring a major 
catastrophe, over the long run planned and actual production will be rea- 
sonably close. 

2. The probabilistic model, on the other hand, contains explicit recognition 
of uncertainty. Such models are of great use in the analysis of advertising 
problems, where the unpredictability of consumers plays a great role. And as 
exhibit ii indicates, they make extensive use of the highly developed theory of 
probability, which has come to be of such great value in the physical sciences. 
One customarily thinks of a physicist as dealing with rather exact concepts 
and highly predictable experiments. Yet physicists faced a problem equivalent 
to the advertising problem in predicting atomic activity. Methods developed 
for physical problems involving mass behavior under random conditions can 
be applied with great facility and value to operations. 

The model is a major goal of the operations research analyst. In one 
sense, the construction of the model, or a faithful representation of the 
operation, is the scientist's primary job. In doing it he develops a theory 

292 Research and Development 

to explain the observed characteristics of the operation. In exhibit hi, 
for example, note how the investigators linked together the salient charac- 
teristics of such diverse and complicated operations as sales, promotion, 
manufacturing, and distribution. The remaining task is to interpret this 
theory through the manipulation of the model, whether mathematical or 


Related to the concept of a model or theory of operation is the measure 
of effectiveness, whereby the extent to which the operation is attaining its 
goal can be explicitly determined. One common over-all measure of effec- 
tiveness in industrial operations is return on investment; another is net 
dollar profit. Measures of effectiveness down the scale might be the num- 
ber of customers serviced per hour, the ratio of productive to total hours 
of a machine operation, etc. 

A consistent statement of the fundamental goals of the operation is 
essential to the mathematical logic of the model. (It does not matter if the 
goals are complex.) Just as the model cannot make 2 and 2 add up to 5, 
so it is impossible to relate fundamentally inconsistent objectives and pro- 
duce consistent and meaningful results. 


The Acme Products Company wanted to set time standards for cost ac- 
counting and labor control on the operations of a battery of taping machines. 
These machines wind a variety of protective tapes on steel cables. The cable 
is pulled through the center of a rotating disk, the "taping head," which car- 
ries a roll of tape; and this tape is unwound through a set of rollers and pre- 
sented to the cable at an angle. Several kinds of metallic, paper, cloth, and 
rubberized tapes are used, and the diameter of the cable treated varies widely. 

The machines used by the company had been purchased at different times 
and were felt to be rather varied in operating characteristics, although the 
principle of operation was the same in all. Time-study methods had failed to 
yield adequate standards because of the complexity and variability of the 
operators' tasks and the uncertain effects of changes in materials. Statistical 
(correlation) methods applied to job records of the time and character of jobs 
failed to explain the variations in time required, and there appeared to be sub- 
stantial differences in efficiency among machines and operators. 

Discussions with operators and foremen indicated that setup and starting 
time and complexity were largely the same for all jobs, but the workers set 
the machine speeds from experience and "feel" of what the tapes used would 
stand without undue breakage. Investigation indicated that the tension in the 
tape was proportional to its speed and the tensile strength was proportional to 
its width. 

The simplest unit of production is the amount produced by the machine in 
one revolution of the taping head, the unit recorded on the work sheets as the 

Operations Research 


"lay" of the tape. If the taping head turns at n revolutions per minute, the time 
required for a job is 


T = n + — 

where U is setup time, / is the "lay," and L is the length of cable in the job. 
From the geometrical relationship shown in Figure 1 the velocity of the tape 

Figure i 

is n\/Tr 2 d 2 + I 2 and the maximum tension the tape will stand is Kn\Ar 2 d 2 + I 2 
where K depends on the strength of tape used. The maximum speed of the 
machine is 


Vtt 2 ^ 2 + I 2 

where Qt depends on the tensile strength of tape material. 

The cable diameter and lay are set independently and an appropriate tape 
width chosen. The required width, from Figure 1, is 


w = I sin 8 = I 

yV 2 ^ 2 + I 2 
and thus the maximum speed for the machine is 


n — ■ 

7r 2 i 2 + I 2 

The time required to cover a cable of diameter d and length L, with a tape of 
material type i at a lay /, is 


T=t + 

Application of this formula to routine job-production records, with appropri- 
ate allowances for the types of tape material used, showed that the operation 
was surprisingly uniform and that the behavior of machines and operators was 
surprisingly similar. Apparent differences were due to unnoticed effects of 
differences in jobs handled. A direct basis was available for setting uniform 
and reasonable time standards. 

. (ir 2 d 2 + / 2 \ 

Kl\ Ud / 


Research and Development 


The classical newsboy problem discussed in Morse and Kimball, Methods of 
Operations Research (John Wiley & Sons, Inc., 195 1), illustrates the construc- 
tion of a simple probabilistic model. While the example itself is trivial, it il- 
lustrates how probabilistic considerations affect results. 

A newsboy buys papers at 2 cents and sells them at 5 cents; he receives a 1- 
cent allowance on unsold papers. He finds by experience he has 10 customers 
a day, appearing at random; that is, he has no regular customers and one per- 
son passing is as likely to buy as the next. Under these circumstances, the Pois- 
son law may be expected to describe the number of customers arriving. The 
chance that m customers will arrive on a given day is given by 




Suppose the newsboy buys k papers and m customers appear. If m is equal 
to or less than k, m papers are sold at a total profit of *pn — k; if m is greater 
than k, k papers are sold at a total profit of 3^. The newsboy's expected profit 

E h = 2 (4m - k)P(m) + 2° 3 &P(ra) 

m = m = k + 1 

The chance he will be able to service all the customers who will pass by is 

S k = 2 P(m) 

m - 

Figure 2 shows how the newsboy's profit depends on the number of papers 
he buys, k, and Figure 3 how his ability to service customers depends on the 
papers he takes. 

Profit: 30< 

Ek 25* 


x x x x v x 













1 1 1 1 1 1 


4 6 8 10 12 14 16 
Papers Taken: k 


Operations Research 



2 4 6 8 10 12 14 
Papers Taken: k 

Figure 3 

Because of chance variations in the number of customers available, if the 
newsboy buys 10 papers every day, he will not average the expected 3 cents 
per paper or 30 cents per day and will meet the available demand less than 
60% of the time. In fact, he can make a little more profit by buying not 10 
but 12 papers daily; if he buys 15 daily, he can make the same total profit on 
the average while meeting the available demand 95 % of the time. 


The Omega Machine Company sells portable industrial machinery, parts, 
and supplies to a large number of industrial users. Orders are received through 
a number of branch offices around the country, with missionary salesmen used 
to visit accounts to explain and promote use of the company's products. The 
company had tried to hold missionary sales expense to a fixed percentage of 
sales, with about 40% of the accounts receiving calls in any one quarter. Statis- 
tical methods based on previous business were used for selecting the accounts 
to be promoted. 

Investigation showed the missionary sales expense to be proportional to the 
number of calls made rather than to the size of accounts called on, and because 
of the nature of the products, total volume from an account depended on the 
number of orders placed, since average order value was essentially the same 
from all accounts. Study of individual account records showed that if pro- 
moted in any quarter an individual account has a probability P(n) of placing 
n orders in that quarter: 

P(n)=— r 


where c is the account's "ordering characteristic." While the ordering char- 
acteristics of individual accounts were unknown, mathematical analysis of sales 

2Q 6 Research and Development 

records indicated that the fraction Q(c) of accounts with an ordering char- 
acteristic equal or greater than c is 

Q(c) = e~ c/s 

where S is the average for the group of accounts. While the value of S varied 
somewhat from region to region and from year to year, the mathematical form 
of Q(c) is remarkably constant. This knowledge permitted comparison of re- 
sults in widely different regions and times, which yielded the results that: 

i. The group of accounts picked in any quarter for promotion by the 
company's procedure showed a distribution of ordering characteristics, c, of 
the form 

Q'AO=^— s 

where g — a/1 — a and a is the size of the fraction selected. 

2. If an account was not promoted in any quarter, there was a 30% chance 
it would be completely inactive, but if active, it would order at a rate only 
70% as great as if promoted; the net effect of lack of promotion was a cut 
of 50% in the potential value of the account. 

These two results were summarized in an equation which showed how 
total business, B(a), depended on the amount of missionary effort expended: 

( fc{i - e - g,s )e- c ' 8 fce- (c + g)/s , ) 
B(a) = N ^\J- y 2 dc + .5 I dc j, = 

o o 


, (j + 2 a -a 2 ) 


where N is the total number of accounts, V is the average order value, and Nd 
is the number of accounts selected by the usual means for promotion. The cost 
of the promotional work would be expressible as pNa. 

A detailed investigation of manufacturing and distribution operations re- 
sulted in the conclusion that the total manufacturing costs (including an inter- 
est charge imputed against capital employed in manufacturing and inventories) 
could be expressed simply, as shown in Figure 4. The break in the curve is due 
to the effect on costs of reaching capacity of existing plants, although opera- 
tions are currently well below this level. 






Total Sales: B 
Figure 4 

Operations Research 2 97 

The facts learned about manufacturing and distribution costs could be com- 
bined with the information on sales and promotion to write down an equation 
for the profit P(a) resulting from any given promotion effort 

P(a)= m (i + za-a 2 ) -pNa- {A - bB(a)} 


where m is the average gross profit and the last two terms are the manufactur- 
ing costs incurred when a volume B (a) is produced. Then 

P(a) = (m-b) (i + 24 -a 2 ) - pNa - A 


By the methods of differential calculus, the profit P(a) will be greatest when 
a is chosen so that 

P'(a) = (m - b)NSV(i - a) - pN = o 

or when 

i — 


The arbitrary method of setting the missionary sales budget as a fixed per- 
centage of sales was replaced by a means for relating sales effort directly to 
profits through the impact on sales and manufacturing. The management had a 
numerical basis for increasing missionary sales effort by over 50% with ex- 
pectation of a handsome return. 

Operations research has frequently brought to light inconsistencies in 
company goals. Take production scheduling, for instance. Very often 
its object has been stated as scheduling production to meet sales forecasts 
with minimum production costs, with minimum inventory investment, and 
without customer-service failure. Yet minimizing inventory investment 
typically requires the use of start-and-stop or at best uneven production 
plans, resulting in excessive production costs; and eliminating the risk of 
not being able to ship every customer order immediately requires huge 
inventories, in the face of fluctuating and at least partially unpredictable 

The solution is to combine and sublimate such otherwise inconsistent 
goals to a higher unified and consistent goal. To illustrate: 

The diverse goals of customer service, production economy, and investment 
minimization can be expressed in terms of costs— the cost of inefficient pro- 
duction (hiring, training, overtime, etc.), the cost of investment in inventory 
(the rate of interest the treasurer wishes to charge to conserve his funds or 
perhaps the return on investment which can be earned through alternative uses 
of the available funds), and the cost of inability to meet a customer's demand 

298 Research and Development 

(estimated loss of goodwill and future business). While the latter two costs are 
primarily policy costs, experience has shown that they are sufficiently de- 
terminable and realistic to afford a basis for management decision. 

The three component costs can then be cast in an algebraic equation express- 
ing their interrelationships in terms of total scheduling cost; and the minimum 
total scheduling cost becomes the one, consistent goal. 

Note that, once set up, the algebraic equation can be worked in reverse. 
Thus, the sales manager might be told how much the company can afford to 
pay for an inventory large enough to avoid varying risks of failure to meet 
consumer demand. 

This kind of clarification of goals is particularly important in relating 
subordinate and over-all company goals— as in the case of a department 
run efficiently at the expense of other departments or of a promotion 
budget based on a fixed percentage of sales without regard to the adverse 
effects on manufacturing budgets. 

The statement of a complete and wholly consistent goal of company 
operations must be recognized as an ideal. Business goals are very com- 
plex, and to catch the full flavor of the objectives of an intricate business 
operation in any simple, explicit statement is difficult. Many business goals 
remain, and probably ever will remain, at least in part intangible— e.g., 
efforts to improve employee morale or contribute to the public welfare. 
To that extent, the objective of operations research must be more modest 
than the construction of a complete model and the measurement of the 
extent to which the operation is attaining the complete set of goals estab- 
lished for it. But it still can serve to clarify the interdependency of those 
intangibles with the company goals which in fact are measurable, thus, 
providing a guide to executive decision. 


The third concept inherent in operations research is that of decision and 
decision making. An essential element in all true operations research prob- 
lems is the existence of alternative courses of action, with a choice to be 
made among them; otherwise the study of an operation becomes academic 
or theoretical. This should be clear from the cases already cited. 

In sum, the objective of operations research is to clarify the relation be- 
tween the several courses of action, determine their outcomes, and indicate 
which measures up best in terms of the company goal. But note that, while 
this should be of assistance to the executive in making his decision in- 
telligently, in every case the ultimate responsibility still lies with him. 


The fourth significant concept concerns the role of experimentation. 
Operations research is the application of experimental science to the study 
of operations. The theory, or model, is generally built up from observed 

Operations Research 299 

data or experience, although in some cases the model development may 
depend heavily on external or a priori information. In any event, the 
theory describing the operation must always be verifiable experimentally. 
Two kinds of experiments are important in this connection: 

1. The first kind is designed simply to get information. Thus, it often takes 
the form of an apparently rather impractical test. In one case the operations 
analysts directed advertising toward potential customers the company knew 
were not worth addressing, and refrained from addressing customers the com- 
pany typically sought— and for a very simple reason. There was plenty of evi- 
dence indicating what happened when advertising was directed toward those 
normally addressed but not enough about its effects upon those not normally 
addressed. To evaluate the effectiveness of the advertising, therefore, it was 
necessary to find out what happened to those normally promoted when they 
were not promoted, and what happened to those normally not promoted when 
they were. 

2. The other type of experiment is the critical type; it is designed to test 
the validity of conclusions. Again, what appear to be rather impractical forms 
of experimentation are sometimes used. Thus, in the most sensitive experiments 
of this type, the validity of the theory or model can often be tested most re- 
vealingly in terms of the results of extreme policies rather than in terms of 
the more normal policy likely to be put into practice. 


Now, before going on to discuss in more detail the administrative prob- 
lems and uses of operations research, it may be well to make clear how it 
differs from other services to management. Many of these services have 
been proved of great value to the business community as a result of years 
of successful application to difficult problems. Are there significant differ- 
ences that make it possible for operations research to extend the usefulness 
of these services? Let us examine some of the leading services briefly for 

Statistics. Operations research is frequently confused with statistics, es- 
pecially as applied to the body of specific techniques based upon probability 
theory which has grown up in recent years. This statistical approach originally 
developed in the fields of agriculture and biology but has now been extended 
into such areas as quality control, accounting, consumer sampling, and opinion 

The operations research analyst does use such statistical methods when ap- 
plicable, but he is not restricted to them. Moreover, there is a difference in 
basic point of view. Statistics is concerned primarily with the relations be- 
tween numbers, while operations research is concerned with reaching an 
understanding of the operation— of the underlying physical system which the 
numbers represent. And this may make a significant difference in results as 
well as approach. In a recent advertising study, the operations research team 
found the key to characterizing the way in which the advertising affected con- 
sumers in the results of a series of "split-run" tests. Earlier, these results had 

j QO Research and Development 

been presumed useless after statistical methods such as analysis of variance and 
multiple regression had failed to show meaningful conclusions. 

Accounting. Operations research is also confused sometimes with account- 
ing, particularly with the control aspects of accounting which have developed 
in recent years. In reality there are several differences. One springs from the 
fact that the fundamental and historical purpose of accounting methods has 
been to maintain a record of the financial operations of the company; and this 
is reflected in the training and attitude of many accountants. The growth of 
the accounting function as the interpreter of information for control purposes 
has been a fairly recent development, and the basic methods used and informa- 
tion provided are strongly influenced by the historical accounting purpose. 

Accounting information is one of the principal sources of data to support an 
operations research study. Accounting data, however, require careful interpre- 
tation and organization before they can be used safely and efficiently. Business- 
men tend to forget that accounting costs are definitions derived in the light of 
the fundamental accounting purpose, and sometimes they tend to confuse ac- 
counting figures with "truth." Operations research, using the same raw data, 
may make other definitions which serve the special needs of the particular 
study. One of the great stumbling blocks in the organization and implementa- 
tion of an operations research study is the disentangling from accounting 
records of the costs appropriately defined and truly significant to the problem 
at hand. 

It is true, however, that in the analysis and construction of measures of con- 
trol, the functions of operations research and accounting do tend to overlap. 
Also, the men working in these functions have strong mutual interests. Ac- 
countants have served a useful purpose in bringing the importance of control 
measures to the attention of business management, while operations research 
has shown ability in building new methods for developing and implementing 
these concepts of control. 

Marketing Research. This management service is concerned with gathering 
and analyzing information bearing on marketing problems. Certain marketing 
researchers do go so far that in some instances they are performing services 
akin to operations research, but for the most part they are content to measure 
the market, by the use of questionnaires, interviews, or otherwise, and to 
gather factual data which management can use as it sees fit. 

By contrast, operations research, when applied to marketing problems, seeks 
to gain a greater understanding of the marketing operation rather than of the 
market itself. Thus, it may rely heavily on marketing research sources for 
data; in one retail advertising study, for example, a consumer-interview pro- 
gram was used to obtain information on the frequency with which potential 
customers purchased outside their own towns. But the objective, even in 
quantitative studies, is usually to obtain a fundamental characterization of con- 
sumers for use in the model. Furthermore, much of operations research in 
marketing problems is directed toward clarifying the interdependencies be- 
tween marketing and other company operations. Finally, it draws on a range 
of techniques and analytical methods that are well beyond the scope of the 
usual marketing research. 

Engineering. Again, the boundary between operations research and engineer- 
ing is frequently unclear. Some examples may serve to draw it more defini- 

i. During the last war a great deal of effort went into the improvement of 
the effectiveness and efficiency of depth charges. The objective of engineering 

Operations Research 301 

and physics research was the construction of a depth charge having the strong- 
est explosive power. Operations research, however, was concerned with the 
effective use of the depth charges then available for the purpose of sinking 

2. In a recent industrial situation, the engineering problem was to construct 
a new railway control system which would get control information quickly 
and clearly to the railway engineer. By contrast, the associated operations re- 
search problem was to determine whether increased speed and clarity of con- 
trol information would help the train engineer in his task of getting the train 
to its destination safely and quickly. 

3. More subtle distinctions can be found in the study of equipment that 
tends to break down in operation, such as aircraft or chemical-process equip- 
ment. The engineering problem may be to find out why the equipment breaks 
down and how the breakdowns can be prevented. The operations research as- 
signment is likely to be finding the best way to run the operation in view of 
available information on the relation between breakdown and use. 

Industrial Engineering. Perhaps the most difficult distinction to make is that 
between operations research and modern industrial engineering. The pioneers 
in the field of industrial engineering did work of a character which operations 
research analysts would be proud to claim for their field. 

In modern practice, however, industrial engineers usually apply established 
methodologies to their problems. Moreover, their work is generally restricted 
in scope to manufacturing activities and, in some cases, to distribution opera- 
tions. Equally important, industrial engineering is not commonly characterized 
by the mental discipline and techniques of analysis that are commonly associ- 
ated with the physical scientist; operations research is. 

Perhaps the most significant difference marking off operations research 
from other management services lies in the type of people employed. 
Operations research people are scientists, not experts. Their value is not in 
their knowledge or business experience but rather in their attitude and 
methodology. It is indicative of the influence which the physical sciences 
have exerted on the people in operations research that they have a self- 
conscious concern with concepts and first principles and show a desire to 
generalize from specific examples to all-encompassing theories. 

In any event, the important point is that, far from supplanting or com- 
peting with other management services, operations research has been 
shown by experience to be particularly successful in those areas where 
other services are active and well developed. Indeed, one useful contribu- 
tion of operations research is frequently that of integrating other informa- 
tion, of using the expert opinion and factual data provided by other 
services in an organized, comprehensive, and systematic analysis. A soundly 
organized operations research group should have available the services 
and counsel of experts in these fields for most effective joint attack on 
management problems. For example: 

In the continuing research program of one retail store chain operation, 
marketing research methods are used to provide field observations, opinions, 
and data on the behavior of consumers. 

Q2 Research and Development 

The accounting organization provides information on costs and capital re- 

In the operations research models these data are combined and interpreted to 
yield information on cost control, staff incentives, merchandise policies, and 
credit management. 


The task of establishing operations research in an industrial organization 
may be broken down into the problems of choosing the initial area for 
investigation, selecting personnel to conduct the work, and developing 
organizational plans for future growth. These problems raise many difficult 
and important questions to which there are no sound answers applicable to 
all companies, but which must be answered by each firm with particular 
reference to its own circumstances and needs. However, certain helpful 
suggestions may be drawn from experience in industrial operations re- 
search to date. 


There are two kinds of starting places: (a) trouble spots where con- 
ventional techniques have failed and management feels the need for ad- 
ditional help and a fresh attack; (b) areas deliberately chosen to test the 
value of operations research because of its possible contribution to the 
general success of the company, i.e., without particular reference to an 
immediately pressing problem. 

There are certain common characteristics of suitable problems on which 
to begin operations research: 

i. There should be an opportunity for decision between alternative courses 
of action. 

2. There should be a real possibility for quantitative study and measure- 
ment. Thus, a preliminary study to provide bases for predicting the acceptance 
of fabric styles had to be quickly dropped in one case because of the inability 
to construct within a reasonable period an adequate quantitative description of 
the complexities of fabric, style, pattern, and color. 

3. It should be possible to collect data. In one case, analysis of accounts 
receivable for the previous two years yielded the key to a knotty marketing 
problem. But, in another case, a study of maintenance problems was found to 
be uneconomical because of the lack of available records showing mainte- 
nance and breakdown histories on equipment. 

4. It should be possible to evaluate results readily. In other words, the 
problem should not be so large that it is indefinite; there should be some 
specific aspect which lends itself to solution. Neither the analyst nor the 
most enthusiastic executive can expect operations research activities to be 
supported on the basis of faith alone. 

Operations Research 303 

The final choice is best made in cooperation with the research team. 
Executives have found it useful to map out the general area in advance; 
the research group can then comment on those aspects which are most 
amenable to study, to clear formulation of the problem, and to likelihood 
of progress with reasonable effort. On this basis a specific problem can 
be selected which meets the requirements both of the executive (for im- 
portance and use) and of the research group (for suitability of existing 
data for quantitative study). 

Much frustration and dissatisfaction can be avoided when the research 
team and the executives keep in mind each others' needs. The research 
team must formulate a sufficiently understandable statement of the prob- 
lem and method of attack to provide the executives with confidence in 
giving support. The executives, in turn, must recognize that in research 
advance specifications for a detailed program including scope and goals are 
frequently difficult and usually meaningless; they must provide the group 
with access to the necessary data and people; and they must maintain 
contact with the work, guiding and redirecting it along the lines of great- 
est value as it develops. 


In perspective, what is the current status of operations research? What 
are its contributions, its limitations, its future? 


Case histories show that operations research provides a basis for arriving 
at an integrated and objective analysis of operating problems. Character- 
istically, operations research tends to force an expansion in viewpoint and 
a more critical, questioning attitude. It also stimulates objective thinking, 
partly because it emphasizes broad purposes and partly because the mathe- 
matical nature of the model and techniques limits the influence of personal 

The results of operations research studies are quantitative. They pro- 
vide an opportunity for sound estimates in terms of requirements, objec- 
tives, and goals, and a basis for more precise planning and decision making. 

The contributions of operations research to business analysis and plan- 
ning have been important and substantial. Here are two worth singling 

1. The application of organized thinking to data already existing within 
the company— Frequently a major contribution has been the location, collec- 
tion, and classification of existing data scattered through widely separated 
branches of the company. In one recent study, an operations research team 
found the same fundamental problem cropping up under various guises in a 

304 Research and Development 

number of different parts of the company. Each division or section had its 
own point of view toward the problem, and each had significant information 
bearing on it that was unavailable to the others. This sort of thing happens 
despite the most sound and progressive management; operations research 
tends to rectify it. 

2. The introduction of new concepts and new methods of analysis— Some 
of these concepts, such as information theory, control theory, and certain 
aspects of statistical mechanics have been carried over from other fields; the 
physical sciences, and in particular modern physics, have been a very fruit- 
ful source of transplanted analytical techniques. But there are also certain 
original contributions, such as the newborn theories of clerical organization 
and consumer behavior, which suggest the possibility of developing further 
tools for attacking important business problems. All these techniques make 
it possible to explore the effects of alternate courses of action before man- 
agement becomes committed to one of them. 


Operations research is hardly a cure-all for every business ill; neither is 
it a source of automatic decisions. It is limited to the study of tangible, 
measurable factors. The many important factors affecting business de- 
cisions that remain intangible or qualitative must continue to be evaluated 
on the basis of executive judgment and intuition. Often they make it 
necessary to adjust or modify the conclusions drawn from the quantita- 
tive analysis of the researchers. Professional personnel in operations 
research strongly emphasize this distinction between the operations re- 
search responsibility for analysis and the executive responsibility for de- 
cision. They point with approval to cases like this one: 

In a recent series of conferences called to implement the results of a long 
and major operations research investigation, the analysts emphasized that 
their conclusions were based in part on the assumption that the output of a 
plant in question could be increased substantially at the existing level of 
efficiency. The executive responsible for the operation of the plant felt 
that this assumption was a sound one. The official responsible for the ulti- 
mate decision, however, decided to follow a more conservative course of 
action than the one indicated by the study, primarily because of his estimate 
of the psychological effect that increases in volume would have on the plant 

The fact that operations research is scientific in character rather than 
expert means that more time is required to achieve useful conclusions than 
in the case of normal engineering analyses. As an applied science, the 
work is torn between two objectives: as "applied" it strives for practical 
and useful work; as "science" it seeks increasing understanding of the 
basic operation, even when the usefulness of this information is not im- 
mediately clear. The executive who plans to support research work of 
this character must be fairly warned of the need for restraint. The natural 
tendency to require that the studies or analyses be "practical" can, if 

Operations Research 3°5 

enforced too rigidly, result in the loss of substantial benefits. Also, the 
results of studies of this type are necessarily somewhat speculative. When 
operations research is purchased, neither the specific program to be 
followed, the precise questions to be answered, nor the successful achieve- 
ment of results can be guaranteed. 

Recognition of this difference between operations research and more 
conventional engineering methods is essential to the satisfaction of both 
the controlling executive and the analyst. 


In conclusion, the future of operations research appears reasonably 
bright at the present time. Successful applications in industry are fulfilling 
the hopes of its early supporters, and the skepticism of businessmen is 
tending to break down as successful case histories pile up and become 
available for publication. 

The areas of potential application of operations research appear broad. 
The future holds possible extensions such as the development of strategic 
concepts through the applications of the much heralded (but as yet largely 
untested) theory of games and by the development of a fundamental 
understanding of the impact of advertising and merchandising methods. 

How will operations research help in the future to clarify the role of 
the executive? Present indications are that it will live up to its expectations 
of helping executives to make decisions more intelligently, but the de- 
cisions will always remain to be made. The possibility of removing all 
subjective and qualitative factors must be deemed at the present time to 
be more a hope than a real possibility, and the construction of completely 
consistent and logical goals, while a reasonable objective in decision mak- 
ing, is probably unattainable. The balancing of the responsibilities to 
society, consumers, owners, and employees will therefore still be the 
fundamental task of executives. 




Donald J. Coppotelli* 

This article describes the integration of substantially all of a large de- 
partment's data into a data processing system kept on a large-scale elec- 
tronic computer. The information thus made available and the advantages 
thereof are discussed. 

In April of 1957 a computer operator pressed the start button of the 
IBM 702 large-scale electronic computer located in General Electric's 
Computer Center in Schenectady and the machine began processing the 
complete financial data for one of the company's departments located in 
the Schenectady plant. This was the culmination of nine months of in- 
tensive research, planning and programming by a small task force of pro- 
cedures specialists and computer programmers. This highly integrated 
system is the first of its kind in General Electric Company although, for 
some time prior to its inception, the computer had been used to produce 
payrolls and other financial data. Although a small department was used 

• From National Association of Accountants Bulletin, XXXIX, 1 (1957), 55-61. Re- 
printed by permission of the National Association of Accountants. 


Departmental Results— from Computer 3°7 

as a prototype for this very practical application of integration through 
electronic data processing, it included most of the elements and financial 
problems of the usual operating department. Complete processing of data 
under the financial data processing system is accomplished by a single run 
on the computer each week and a separate monthly closing run. 

As an example of the scope of the data being processed through the in- 
tegrated system, the department using the system includes sixteen major 
components and more than one hundred reporting units (line foreman 














Traveling & 
Living Expenses 


Internal Stock 



Stock Withdrawal 


Accruals & 


Special Transfers 




Manual Over- 
head Adjustment 



Tape 1 

•-Input to weekly financial data processing run 
••-Transaction Codes 

and first line supervisors) each requiring specific reporting to maintain 
control at the workers' level. Accounting records are maintained for more 
than eight thousand individual shop orders and accounts. An average 
week requires the processing of more than four thousand labor vouchers 
and seven thousand other financial transactions. Shop workers are on 
hourly day-work. Professional and office employees are paid on the basis 
of weekly or monthly salary. All time is distributed weekly. Costs are 
maintained on a job order cost system. Prior to setting up the integrated 
system, punched card equipment was used to perform most of the ac- 
counting work of the department. 

308 The Production Process and the Administrator 

The study team developed ten basic concepts around which the system was 

i. Shop order master file on magnetic tape 

2. Automatic payments to vendors 

3 . Weekly shop order status reporting 

4. Weekly voucher comparison reports 

5. Mechanized sub-ledgers 

6. Automatic monthly billing 

7. Mechanized overhead reports 

8. Early monthly closing 

9. Integrated multi-purpose registers 

10. Complete accountability at the foreman and line-supervisor level 

One other feature was considered by the study team but was not in- 
corporated in the system. This was the maintenance of inventory and in- 
ventory control records by the computer. However, this particular 
department's inventory was relatively small and therefore was not incor- 
porated as a part of the system. It could easily have been included and, in 
other departments establishing similar type data processing system, prob- 
ably will be. 


The shop order master file on magnetic tape is the basis of the financial 
data processing system. Inasmuch as central compact storage of complete 
financial data is the underlying principle of the integrated computer sys- 
tem, a basic understanding of the shop order master file is necessary to an 
understanding of the complete system. This file contains over eight thou- 
sand individual records stored on a single reel of magnetic tape. It con- 
tains a record of each shop order, sub-ledger account and balance sheet 
control account in the department. The records are maintained on the tape 
in a sequence which permits weekly and monthly reports to be extracted 
as a finished product, with no sorting required. This is an important con- 
sideration in any tape-operated computer, since sorting is one of the least 
economical operations to perform on large machines of this type. 

Each individual unit record in the system contains a number of fields. 
Each record contains the same fields and these fields are consistent in 
length and content, even though some of them may not be used for all 
types of records included in the system. These fields fall into two general 
type categories: indicative data and statistical data. Information stored in 
the indicative data fields includes such items as requisition number, shop 
order number, appropriation or cost estimate number, shop order title or 
description, and budget data. Also included in the indicative fields are a 

Departmental Results— from Computer 3°9 

number of codes which the computer will interrogate when processing 
the record to determine which overhead rate should be applied to direct 
labor from a table of overhead rates stored in its memory. Statistical fields 
include current-month accumulations of charges by classification of the 
charge. For example, material, various classifications of labor, overhead 
and billing liquidations are each accumulated separately. 

Year-to-date amounts and total-to-date are also accumulated by the 
same classification of charges. Unbilled balances by billing classification 
are also maintained for each shop order and account. As a result, all data 
required for general accounting, cost accounting and some elements of 
purchasing and orders-received accounting are included in the master file 
tapes. All registers, reports, billing and control totals are extracted from 
data summarized in the master tape on a weekly, monthly or annual basis. 


Monthly journal entries have been replaced by transaction records in 
the form of punched cards which are prepared currently for all financial 
transactions. This has completely eliminated the journal entry system of 
accounting in this department, since it is no longer necessary to journalize 
charges in order to post to the individual records. The computer posts all 
details affecting an individual record at the same time, instead of accu- 
mulating all charges of a particular type in journal entries for posting to the 
records affected. Transaction code numbers replace the former journal 
entry numbers but accomplish considerably more than journal entries did 
under manual or mechanical systems. 

exhibit i shows the manner in which transactions for a given week are 
converted to tape for processing by the computer. It will be noticed that 
file maintenance type changes bear the lowest transaction numbers. Since 
weekly transactions are put on tape in transaction number order, this 
means that any file maintenance changes, such as the opening or closing of 
a shop order, will enter the computer before any changes occur in statisti- 
cal data. For example, a shop order may be opened in a given week and 
charges booked to it. In addition, the transaction code indicates to the 
computer the exact routine from a library, of routines stored in the com- 
puter's memory, which should be selected for the processing of a trans- 
action. The transaction code tells the computer what changes need be 
made in the shop order master file or updates the purchase order master 
file or specifies the reports into which the detail card must be incorporated 
or establishes any special accumulations which are to be accomplished for 
control purposes. 

3 io 

The Production Process and the Administrator 



At the end of each week, the transaction cards are converted to a 
weekly transaction tape which, along with the weekly labor distribution 
obtained as a by-product of a prior payroll run, is used to update the 
thousand records contained in the shop order master file and, at the same 
time, produce current weekly cost and control reports. This is illustrated 
by exhibit 2. The exception principle is followed in determining which 
shop orders and accounts are to be reported in any week. If a manager, 
supervisor or foreman does not desire a report on a specific shop order, he 
so indicates and this shop order account is not reported to him until he 
asks that it be included in his report. 

In the purchasing department, as each order is placed a by-product 
punched card is automatically produced. Information in this card includes 
the shop order number, purchase order number, and dollar value of the 
order. These cards are converted to a purchase order master file on mag- 
netic tape in purchase order number sequence. When the bill is received 
for payment, the purchase order number and the gross amount of the bill 

Departmental Results— from Computer 3 1 1 

is converted to magnetic tape and fed into the computer along with the 
purchase order master file. The computer prepares checks and pay war- 
rants covering all purchase orders for which material has been received 
and invoices are on hand. As the computer processes the payments to 
vendors, it automatically calculates the net amount of the bill and the 
amount of the discount. Payment will not be made unless a purchase order 
is available in the purchase order master file. This establishes a check 
against duplicate payments. In addition, if the net amount of the invoice 
exceeds the amount of the order, an automatic message is typed out on 
the typewriter on the computer console to indicate this fact. 

As output from the weekly computer run, a shop order status report is 
prepared by the computer indicating the charges incurred during the week 
by classification of the charge, total-to-date expenditures and, on any ac- 
count with a budget or authorized amount, the unexpended balance is 
printed. A weekly report is prepared for each of the more than one hun- 
dred units on the foreman and line supervision level. 

Each week the computer automatically provides a voucher comparison 
report for both weekly and monthly paid employees as direct output from 
the payroll run. These reports show any excess or shortage of vouchers 
for each employee, both weekly and monthly, based on clock card hours 
or an assumed forty hours if no clock card has been submitted for the 


-Separate Report for each Laboratory Department 

-This SOMF will be used as input to the next weekly financial data processing run. 

3 1 2 The Production Process and the Administrator 

week. Vouchers are not required for any employee whose time is normally 
charged to a standard overhead order, since the computer automatically 
distributes this time to the standard expense account in the employee mas- 
ter file. 


One of the major advantages of the integrated system has been the early 
closing of the department's books. On the first Tuesday following the 
close of a fiscal month, the books are completely closed by a special run 
on the computer. This is illustrated by exhibit 3. This run prepares a trial 
balance, a general ledger, overhead reports and an orders received report 
by each component of the department. All charges, including labor in- 
curred in the last week of the month, are on an actual basis and are not 
accrued. Therefore, the trial balances and sub-ledgers include all of the 
charges distributed to several thousand shop orders. Under this system, if 
desired, the books can be completely closed weekly. During the conver- 
sion months of April, a weekly trial balance was actually prepared by the 
computer to localize any possible errors made as a result of unfamiliarity 
with the new system. 

A monthly sub-ledger register is produced from magnetic tape, which 
includes all of the sub-ledgers supporting general ledger accounts. These 
include plant equipment, engineering orders, raw and in process inven- 
tory, overhead, plus several that are special for the department. This 
register eliminates the need for any manual posting formerly required. 
During the monthly closing run, all charges incurred against orders which 
may be billed during the current month are automatically billed out. In- 
voices are automatically prepared by the computer, including appropriate 
classification of charges for statistical purposes. Most of the billing of this 
department is done internally to other departments within the company 
and the computer, in addition to providing the bills, also provides billing 
schedules by company destination. The integrated system results in the 
issuance of billing two to three weeks sooner than had been done pre- 

During the closing run, an overhead report is automatically prepared 
for each component in the department. These reports show expenditures 
against each overhead account on a current month and year-to-date basis. 
In addition to expenditures, the overhead report also contains budget 
analysis. The computer calculates the variable budget based upon adjusted 
applied labor for the month. Performance against the calculated budget 
is shown as a percentage. Audit symbols appear opposite the overhead 
account when expenditures exceed the calculated variable budget amount. 

Future of a Successful Product 3 1 3 

The financial data processing system has fulfilled all of the basic ob- 
jectives set for it and is running successfully on the computer. It has per- 
mitted much earlier closing of the department's books, more current bill- 
ing of the department's books, more current billing of the department's 
expenses, much earlier reporting to management and, thus, appreciably 
better management control through financial operation. It has resulted in 
an appreciable saving in clerical and machine accounting costs. 



J. Curran Freeman* 

This article tells the story of the planning of expansion related to a spe- 
cific company. It covers the planning of sales revenue, costs, profits, and 
return on investment and involves facilities and working capital planning. 

There is sufficient evidence to convince even the most conservative of 
us that we stand on the threshold of a "golden age" in the economic and 
industrial life of America. We shall participate in the fruits of this great 
national growth directly in proportion to the thoroughness with which 
our planning is done. We need to undertake intelligent planning, based 
on supported facts extrapolated on the basis of careful projections and 
flexibly designed for quick modification as circumstances become more 
proximately known and evaluated. 

This paper will not treat financial planning from the viewpoint of cur- 
rent interest in the subject of raising new capital to meet the needs of the 
greatest capital expansion in history. Provision of capital is, or may be, 
periodic. Hence discussion of its effective use, after it has been provided 
should, in my opinion, be of greater interest. This article affords us the 
opportunity to 

i. Review a product (not an end-product) projection study just com- 
pleted for one of our companies, commencing with the market analysis and 
culminating in a profit and loss projection. . 

2. Explain briefly our approach, on a project basis, to manufacturing and 
engineering problems associated directly or indirectly with the expansion. 

3. Accord proper recognition to organization structure. 

4. Highlight our control accounting techniques. 

*From National Association of Cost Accountants Bulletin, XXXVIII, 9 (1957). 
1 099- 1 105. Reprinted by permission of the National Association of Accountants. 

3 14 The Production Process and the Administrator 

The study described was essentially the work of a team. The staff di- 
rector of marketing was responsible for directing the work and for final 
consolidation and drafting of the study. He was aided by line management, 
executive as well as sales, finance, manufacturing and engineering. 


The history of the product has been one basically of the inability of 
manufacturing facilities to keep pace with sales potential. Consequently, 








Number Un 

its Application 





or New 




Produced in Proc 












X X 
















X X 






































X X 







* Original Equipment Market 

it was necessary to do a market research study to find that potential. The 
product has an application in many markets. Hence each market in which 
the product has potential application was analysed in terms of units pro- 
duced in 1955. These end products were then identified as to whether 
they might (possible applications) or would not (negative applications) 
provide a market for the product under review, (exhibit i) The number 
of end products extended by the units of our product developed its po- 
tential in terms of units and dollars for the period under projection, 1957 

Future of a Successful Product 3 1 5 







$ 796,000 






Total Total 
Potential Potential 

32.2% $ 862,000 

20.0 613,650 

8.2 1 131693 

24.0 3,299,000 


% of 









$ 4 2 7^2 




1 00.0 

* Columns for 1958 and 1959 omitted 

through i960, as shown in exhibit 2 which also carries figures for Dresser 
historical sales and percentage participation in the potential market. Pro- 
jections were extended to possibilities of further market penetration based 
on price, quality and sales coverage. Through field study and customer 
contact, potential was verified by a study of competition, whether direct, 
competitive or comparable method. It can be appreciated that findings of 
the field study have multiple value in research and product engineering as 
well as manufacturing techniques and costs. To finalize the program in 
terms of production, it was also necessary to make a detailed study of dis- 
tribution methods and techniques, not previously required in a market the 
potential of which so outstripped production capacity. Anticipating the 
results of this study in terms of forecast sales, for purposes of perspective 
let us say at this point that the study projects a multi-million dollar in- 
crease in sales of 300 per cent in i960 over 1955. 


Thus far we have reviewed the market potential and product forecast 
approach, the latter in terms of unit and dollars. Looking to production 
problems, the study proceeded from this point to project units and dollars 
in terms of styles and sizes, total potential as well as estimated Dresser par- 
ticipation. This detail is neither academic nor sophisticated, if the practical 
problems of manufacturing conversion (men, methods and machines) are 
kept in mind. It provided the starting point for manufacturing projection. 
The analysis of the manufacturing aspect of this expansion covered: 

1. Equipment requirements and production capacity 195 5- 1960 (Equip- 
ment investment). 

3 1 6 The Production Process and the Administrator 

2. Equipment (selection and automation). 

3. Manpower requirements. 

4. Plant layout and plant location. 

5. Manufacturing expense. 

6. Cost reduction. 

a. Material cost. 

b. Labor cost. 

7. Detailed manufacturing problems to be resolved. 

Following the bill of materials "explosion" techniques, units of the 
product were reduced to parts requirements in terms of present equip- 
ment capacity and projected requirements for machine capacity during 
each of the years under study through i960. As shown in exhibits 3 and 
4, capacity requirements were developed by parts total, machine, and 
capital expenditure in amount and cash date. Summarized in terms of in- 
dividual machines and dollars, a proposed equipment investment of over 
$2,000,000 became evident. The equipment selected for immediate pro- 
curement is the net result of a complete evaluation of a two-year study 
and continuing conference with machinery builders, including automation 
studies. Based upon the production requirements as indicated through 
i960, and the inventory of equipment required to effect this production, 
a man-power study was completed in the detail of machines, shifts and 
hands required. Understandably, expanding a business 300% in five years 



(Columns for 1957 and 1959 omitted) 


Part A (Prod. Required) 
1956 Present Equip. Capac. 
* ( 1 ) New Equip. Type 


(2) " 

Part B (Prod. Required) 
Present Equip. Capac. 
( 1 ) New Equipment 







*i 95 6 



1 1 ,800,000 



Assembled Product (prod.) 
Present Equip. (Capac. O.) 

* on order 



Future of a Successful Product 


also required a plant layout study. This study included plant location and 
activity relocation covering the usual items of freight rates, labor availa- 
bility and labor rates, detail layout of manufacturing and service facilities, 
and total expenditure required. Completing the preparation of raw data, a 
detailed yearly manufacturing expense projection was prepared through 

exhibit 4 

Product XYZ 







1956 DATE 




80,000 1956 

80,000 1957 

160,000 1958 



















1 80,000 



Having the above data in hand we were able to complete a projected 
profit and loss statement in the customary manner. The treatment of wage 
differentials (on a physical relocated plant basis) and manufacturing cost 
savings deserves a word of comment. Labor rate differentials, particularly 
on a union- non-union basis, were not allowed as a cost saving claim. Time 
can conceivably wipe out the differential and we believe that long-range 
decisions to invest should be supported by long-range considerations. 
Though there may be good and sound arguments for estimating manu- 
facturing cost savings, we accepted the program strictly on a straight 
dollar profit increase, producing the same percentage to sales as presently 
generated. These statements apply to the long-range. For the fiscal period 


The Production Process and the Administrator 

immediately ahead cost savings claimed are critically reviewed and evalu- 
ated cost- and profit-wise. 

To complete the financial projection, estimated balance sheets through 
1 960 were prepared to develop working capital requirements and the base 
for the return on investment calculation. It is interesting to note from ex- 
hibit 5 on which relevant data is graphically presented that return on in- 
vestment declines rather sharply to 1959 and begins the upturn at that time. 
Long-range planning has many advantages. Projected far enough, the re- 
turn on investment sometimes becomes a deciding factor to invest where 
short-range planning or rule-of-thumb fast payouts might cause an overly 
conservative management to forego a profitable opportunity. 

exhibit 5 










A. $ XXX 

B. $ XXX 

C. $ XXX 














(Before Taxes) 


A. $ XXX 

B. $ XXX 

C. $ XXX 


A. $ XXX 

B. $ XXX 

C. $ XXX 


55 19 

56 19 

57 19 

58 19 

59 19 


Future of a Successful Product 



Facilities, whether land, building, or equipment, are at best inanimate 
tools which require management activation to produce effectively and at 
a profit. It is a truism to say that what is every one's responsibility is, in 
the last analysis, no one's responsibility. We believe in clear cut delegation 
of authorities and responsibilities. In the instance of this expansion, sub- 
stantial as it will be, the exigencies of the business require progressive de- 
centralization of profit responsibility, eventuating in a general managership 













as indicated in the Organization Plan (exhibit 6). In its initial stages we 
contemplate a product manager with complete profit planning responsibil- 
ities on a profit center basis, encompassing: 


New designs for existing markets. 
New designs for new markets. 
New markets for existing designs. 

Putting Plans Into Effect 

In all areas of 
Product research. 

Manufacturing— Tools, inventories, etc. 
Purchasing— Sources of supply. 
Cost reduction. 

Review and Control 

Responsibility for results (Profits) 

320 The Production Process and the Administrator 


We have now reviewed product planning, set the plan of operation, and 
arrived at the point at which there is necessity for a system of review of 
actual results against plan as a basis for remedial action, where necessary, 
to control according to plan. In my experience, I have found the relation- 
ship of operating profit to gross assets to be the best measure of operating 
efficiency. Operating profit is the result of sales revenue less manufactur- 
ing costs and operating expenses (sales, engineering, administrative and 
general expenses). To be sure, there are other bases to which to relate this 
figure, such as net invested capital or net assets. However, gross assets as a 
ratio denominator measures a business on the basis of continuance instead 
of wasting of assets. We consider the use of fixed and variable cost and 
expense elements as well as the direct costing of product (adding a com- 
plement of fixed expense for inventory valuation, however), as being the 
most effective means of developing "responsibility costs" whether at a 
profit center or cost center level of management. Finally, where control of 
cost cannot be readily measured in terms of dollars and cents, such as 
broad investigative and research programs in manufacturing, engineering 
and sales, a project control is used. In project control we set the objective, 
the calculated methods of resolution, the time-phasing, and progress re- 
porting routine. 


Long-range planning as a "new dimension" in industry (Ralph J. Cordi- 
ner before the Economic Club of New York, March 1956), means depth 
of penetration in terms of investigation and perception. Depth of percep- 
tion can come only from intensive planning born of critical, painstaking 
and minute analysis. This is the route we have taken in Dresser and have 
importantly changed the character of the company's operations in recent 
years. As recently as 1949, approximately 90 per cent of its sales volume 
consisted of heavy capital equipment having a long service life. As a result, 
sales and earning of the company, as then constituted, tended to exhibit the 
fairly substantial fluctuations typical of the capital goods industry. Subse- 
quent acquisitions, as well as internal growth of the Dresser Companies, 
have been directed primarily toward increasing the portion of the total 
sales represented by expendable products and technical services. This plan- 
ning has paid off handsomely. In the current year expendables will ac- 
count for 60 per cent of our sales. While accomplishing this basic change, 
we have advanced detail profit planning from one year planning (fiscal 
year) to five years and now to ten. 

"Clocks" for Management Control 321 

That a long-range plan, even a five-year plan, may be subject to change 
causes no concern. Operating in a dynamic economy, we expect that to 
happen and, consequently, have, as management, designed all our plans, 
whether for material, men, money, or methods to allow us to quickly 
adjust to new found circumstances. 


Allen W. Rucker* 

The author develops the concepts of break-even point control into a 
helpful system which can help the executive to control his over-all business 
costs and maximize his earnings. 

The top executive in modern industry wants to know both profit-wise 
and cost-wise "where he is as against where he ought to be" at any given 
time. He wants to make planning and policy decisions knowing what the 
results must be in order to stay within certain profit and cost standards. 
He wants to be able to "manage by exception." He wants to know what 
his associates are accomplishing— and to let them know that he knows— in 
concrete dollar terms; for then praise and blame can be phrased in objec- 
tive quantitative terms so as to spur constructive achievement. 

How can the executive do these things? One of the most helpful means 
is breakeven point control. Now, this is not a new concept. Early this cen- 
tury C. E. Knoeppel, often called the "father" of breakeven point control, 
developed the fundamental distinction between fixed (better, rigid) ex- 
penses and variable costs contingent on volume. 1 But thinking about the 
concepts has not been static in the intervening years. There have been 
many discussions and publications on the subject. In recent years, in par- 
ticular, a number of progressive companies have been working with break- 
even point control, adding refinements to it and developing its practical 
effectiveness. As a result, the concept has been carried well beyond the 
"textbook" version which businessmen can find in libraries. 

* From Harvard Business Review, XXXIII, 5 (1955), 68-80. Reprinted by permis- 
sion of the Harvard Business Review. 

1 C. E. Knoeppel, a series of articles appearing in The Engineering Magazine, Vol. 
XXXVI, October 1908-March 1909; published in book form under the title Graphic 
Production Control (New York, 1918). 

322 The Production Process and the Administrator 

In the pages to follow I shall discuss some applications of the breakeven 
principles which are destined, I believe, for wider acceptance among pro- 
gressive companies in the years ahead. These applications center about 
business "clocks" which can be constructed to help an executive make 
sound decisions on such problems as whether: 

The company will be able to pay an extra dividend if the present rate of 
operations continues. 

The company is expanding faster than the growth of volume justifies. 

The maximum permissible limits on variable selling costs in opening a new 
territory or in selling a new product have been reached. 

A proposed advertising appropriation is sound. 

A change of emphasis in selling effort among different areas is needed. 

A union wage demand is justified. 

The construction of these "clocks" is explained in the Appendix (where 
the various exhibits referred to in the text appear). For the top executive 
it should suffice here to point out that the "clocks" can be built for his use, 
that the making of them need not absorb his time and energies. For him, 
the paramount considerations are those of utilizing the "clocks" to con- 
trol his over-all business costs and to maximize his earnings. 


Modern industry, with its heavy plant investment, depreciation sched- 
ules, taxes, and so on, has probably made the present generation of man- 
agers as "timetable" conscious as almost any other group in the country. 
Company executives are continually concerned with questions pertaining 
to whether operations are proceeding according to schedule or not. 


For example, here is one question commonly asked: "How much vol- 
ume does our firm need in order to pay an extra dividend, how much just 
to cover the regular dividend, how much to break even? " 

The volume spread between the breakeven point and the extra-dividend 
point may be as much as 150%. Or, as one management recently found, an 
insignificant 6 % increase in volume over that forecasted in the company's 
executive-committee budget would mean a jump of 49% in pretax profit. 
That revelation sparked an immediate revision of sales quotas and the 
launching of two new products. 

If a senior executive wants to spur his organization to new accomplish- 
ments, let him picture the possibilities in figures that disclose the end re- 
sults. Not only production managers and sales managers but also directors 
need to have a perspective view of a firm's prospects. In one instance, the 

"Clocks" for Management Control 323 

previous reluctance of the directors to approve an expansion program 
evaporated in five minutes when the board was shown that the resulting 
20% gain in output would raise earnings on net worth from 8.6% to 
12.2 % and lift pretax profits almost 90% above the prevailing level. 

exhibit 1 illustrates such a perspective view. This table, prepared for a 
moderate-size firm, discloses an exciting fact previously hidden even from 
top management: every dollar of volume above the breakeven point gen- 
erates a pretax profit of 33.88 cents. If this carrot is not enough to spur 
action among executives, then the club which the table also provides can 
be wielded-every dollar of volume below the breakeven point will throw 
33.88 cents in red ink on the operating statement. 

(Note that in this and all other exhibits, volume is expressed in terms 
of production value rather than sales, for reasons explained in the Ap- 

Other questions that are asked concern expansion. In management 
circles there is an uneasy awareness that breakeven points have been rising 
under the impetus of monetary inflation as well as with the expansion of 
facilities. I find that a growing number of senior executives are asking: 
"Are we going ahead with plant expansion faster than our realizable in- 
crease in volume? Are we increasing all forms of rigid expense-deprecia- 
tion, maintenance, research, and staff personnel expense-at a rate greater 
than the growth rate of our business justifies? " 

These questions, too, can be answered on a factual basis, using a graphic 
approach such as exhibit ii. This chart, based on the ratio of actual volume 
to breakeven volume over a ten-year period, shows the extent to which the 
volume realized in any year exceeds that required to break even (taken 
as 100%). The five-year or ten-year trend of this ratio-the "safety factor 
index"-is as revealing to management as a temperature chart is to a phy- 

With this type of visual control, a manufacturer of metal products 
maintains an up-to-date comparison of (a) his actual volume and (b) the 
volume required simply to break even. Whenever the chart (the figures 
will do just as well) discloses a tendency for the safety factor to decline, 
this manufacturer knows that expansion of his organization and plant is at 
a rate greater than his realized business growth, to the detriment of profit. 
The reverse is also true; a persistent upward trend of the safety factor in- 
dex may signal the need for expansion. 


Both exhibit 1 and exhibit ii represent two forms of a business "clock" 
showing whether a firm is on time, behind time, or ahead of time, exhibit 
1 clocks business progress during the current fiscal year, showing what 
expense and pretax profit ought to be under any level of business activity; 

324 The Production Process and the Administrator 

exhibit 11 clocks the changing spread between actual and breakeven vol- 
ume. It should be emphasized that exhibit i is not a forecast of business 
conditions, nor of business costs and business profits. On the contrary, this 
table is intended to show what total expenses should be, and what pretax 
profit should be, not only at the forecasted volume but at any other 
volume level above or below the forecast. For example: 

On Line II under Column B appears the company's forecasted or budgeted 
volume for the fiscal year. Directly opposite on the same line, under Column 
C, appears the standard total expense allowed at this volume (the actual com- 
putation is explained in the Appendix)', and in Column D the standard pretax 
profit that would result. Note that the table also discloses the allowable ex- 
pense and standard pretax profit for any volume level above or below that 

In brief, failure to make the expected volume does not leave the executive 
without a profit goal on the lower volume actually realized. So also success in 
surpassing the expected volume does not lull management into complacency 
because profits are larger than budgeted; the table discloses whether or not 
profits are as much greater as they should be for the greater volume actually 

This type of control table not only shows the senior executives, almost 
at a glance, how much expense deviation from budget is permitted for 
any fluctuation in volume above or below budget, but makes it possible 
for executives to distinguish quickly between (a) expenses caused by 
changes in volume and (b) expenses caused by changes in the effectiveness 
of internal control of operations. To spell this out: 

Changes in volume (except when due to newly added facilities) do not auto- 
matically cause a change in rigid expense, but they do cause, almost automati- 
cally, a change in the total amount of budgeted variable contingent costs. In 
exhibit 1, standard changes in these costs are allowed at the rate of 66.12% of 
changes in volume, up or down (the computation of this figure is explained in 
the Appendix). For example, the difference in total costs between lines 5 and 
6 ($243,414) is 66.12% of the difference in volume shown opposite these two 
lines— i.e., 66.12% X $368,147. 

On rising or falling volume, any change greater than this 66.12% is an indi- 
cation that internal management control over costs is less effective than that 
required to meet planned results; any change less than 66.12% shows that 
control is more effective than planned. 

By means of such "clocks" senior executives can more easily call a halt 
to alibis and to rationalizations of a poor showing, or, for that matter, 
of an improved showing that still fails to come up to standard. 


Management will find it convenient to divide the uses of the control 
"clocks" into two categories: (1) business policy and planning decisions; 
and (2) operations control, both of costs and of income. 

"Clocks" for Management Control 325 

A few examples from experience will illustrate both types of use and 
will suggest a number of others. 


Expansion of plant by either new building or acquisition of other firms 
increases both production value and (because of increased depreciation 
and executive supervisory personnel) rigid expense. Hence, the breakeven 
point tends to rise. This is normal. The factor to watch is the ratio be- 
tween rising production value and a rising breakeven point. If this ratio 
shows an upward trend, the expansion policy is justified: if the trend de- 
clines, the management may be overdoing matters. In the case of the com- 
pany concerned in exhibit ii, for instance, it was evident that plant 
growth was tending to outstrip the gain in production value; the safety 
factor signaled "caution" with respect to further additions to capacity. 

Limiting Sales Expense. The experience of a manufacturer of a diversi- 
fied line of printed products illustrates another use of these "clocks": 

In a situation marked by rising sales volume but also by diminishing profits, 
the locus of the trouble was traced to sales expense above standard. A break- 
down of production value by sales areas resulted in a statement for each area 
similar to that in exhibit iv. 

In two geographical areas, it was found that variable contingent sales expense 
not only exceeded the over-all company average but actually exceeded the 
entire operating margin above variable manufacturing and administration ex- 
pense. (In the case of the company in exhibit iv, a comparable situation would 
have existed if item 4-b-2 had been 48.54%.) In those two areas the firm was, 
in effect, giving away the merchandise plus 9%; that is, total variable costs 
were 109% of production value. 

The sales manager was given six months in which he could either correct 
the situation or abandon all sales effort in those areas. 

This is not to say that variable sales costs in some territories may not 
properly be higher than the company average for that type of expense. 
But there is a limit, and it obviously must be somewhere under 100% of 
the production value generated by the territorial sales. With variable con- 
tingent costs once determined, the precise maximum limit on variable 
selling costs in introducing a new product or opening a new sales area 
without a cash loss can be found from a statement like exhibit iv. To 

Production value 100% 

Less: variable manufacturing costs (includes labor's share 

of production, or the cost of hourly rated factory labor) 52% 
Less: variable administration expense _5 57% 

Maximum permissible for variable selling costs without suf- 
fering an out-of-pocket cash loss 43% 


The Production Process and the Administrator 

Naturally, no one deliberately goes up to this limit or above it unless 
he has to, and then only temporarily. But without knowledge of this limit, 
who is to know by how much the limit is exceeded and for how long? A 
"clock" of the type shown in exhibit iv for each prospective new sales 
territory or new product can be invaluable to top management. A like 
treatment may well apply in some or all old territories. 

Exploiting New Markets. The same approach can also be used to dis- 
close whether or not management is too timid in exploiting new sales 
areas and new products. In many firms, management's figures are arranged 
to show the rate of pretax profit by sales territories or by products. Now, 
in each case this figure will always be less than the operating margin, be- 
cause of the practice of allocating existing rigid expense (overhead) on 
some pro-rata basis. Note that in exhibit i, for instance, all the profit rates 
are considerably less than the operating margin, which is 33.88% of pro- 
duction value (see exhibit iv, item 4-b-4). Actually, however, this need 
not be the case. Up to the whole amount of the increase in operating mar- 
gin could be spent if necessary to open a new sales territory or to exploit 
a new product without causing a cash loss and a drain on profits; any ex- 
penditure under this upper limit will add to the firm's pretax profit (in 
dollars, not as a percentage of volume). 

The reason for this is that the old sales areas or old products are al- 
ready carrying the existing rigid expense or overhead; the addition of a 
new market or product does not necessarily add to the already existing 
amount of rigid charges. I realize, of course, that this concept is a heresy 
to the cost accountant, who tends to insist on "total costs" and on allo- 
cating to each new product or new sales territory its share of rigid ex- 
pense. My rebuttal is that this practice overlooks the fact that a shift of 
overhead to new products or territories reduces the burden on the old 
ones and thereby increases the apparent profit on them. What is more, 
it misleads top management. 

Reallocating Sales Effort. Here is another concrete example of what 
can be achieved in profit gain using the same methods: 

A manufacturer of a wide line of textiles sold nationally wanted to discover 
the cause of the firm's inadequate earnings. He attacked the problem by mak- 
ing an analysis of production value by product lines. For each product line, he 
assembled his variable manufacturing, administration, and selling costs in the 
form shown in exhibit iv. 

He then classified the different product lines in two groups: (a) those with 
an operating margin below the company average, and (b) those with an op- 
erating margin higher than the company average. The first group accounted 
for almost two-thirds of total physical volume but for less than one-third of 
total dollar operating margin. 

This situation— not an uncommon one— was clear evidence of misdirection 
of sales emphasis. Accordingly, top management instructed the sales manager 
to shift the emphasis from the high unit-volume products to those with the 

"Clocks" for Management Control 327 

higher-than-average operating margin. In nine months the firm lifted its pre- 
tax profit almost 100%, despite the fact that the physical volume of sales 
slightly declined. 

Advertising Appropriations. In most progressive companies executives 
are continually coming up with new ideas and schemes. How can they be 
evaluated? A breakeven point analysis can be extremely helpful in evalu- 
ating a proposal from a dollars-and-cents point of view. To illustrate, let 
me take another instance involving the company in exhibit iv: 

The sales manager insisted on an additional $100,000 for advertising. He 
argued persuasively, as only sales managers can, that the added appropriations 
"will get us a big increase in volume." 

The president took a look at his firm's operating margin (exhibit iv, item 
4-b~4). Then he divided $100,000 by 33.88% and came up with $295,159 as the 
added production value required simply to recoup the $100,000 of added ex- 
pense, with no profit whatever. In other words, he brought the discussion out 
of the realm of generalities and down to earth. He went further; he added to 
the $100,000 a factor of one-third for pretax profit and again divided by 
33.88%. The resulting figure, $393,545, showed him the gain in production 
value needed to justify an extra $100,000 in advertising without violating the 
company's profit standard. 

"See here, Jim," he said, "what do you mean, 'a big increase in volume?' Are 
you prepared to say you can get as much as $393,545 production value in re- 
turn?" The sales manager had not thought things through. Confronted with 
this concrete figure, he backed down. 

Labor Demands. One of the prime fallacies of these times is the notion 
that "we can raise wages because last year's profits were pretty good." 
The truth is, of course, that there is no way by which last year's profits 
can be shared in next year's wage rates. No set of figures affords a simple 
solution to union wage demands, but with a definite standard of the per- 
missible percentage relationship of wages to production value, any firm 
can quickly find out what it can and cannot do. A problem tends to be 
simpler when reduced to concrete figures. With them, management can 
have a constructive wage policy; it can make sound, rational decisions 
that it can stand on both at the bargaining conference and at the directors' 
meeting as well. To illustrate again from the experience of the company 
in exhibit iv: 

The union demanded five cents an hour plus "fringes," or a total of seven 
cents, as "our share of increased productivity." Before saying either "yes" or 
"no" the management took a look at one of these "clocks." As shown by 
exhibit iv, item 4-a, standard labor cost was 47.61% of production value. The 
current performance was 46.18% and hence allowed for some wage rate in- 
crease. How much? With gross hourly average earnings, including all "fringes," 
at $1.81, the permitted increase was easily found by the formula: 

47.61% (standard) 

X $1.81 

46.18% (actual) 



The Production Process and the Administrator 

The permitted increase without exceeding the standard was thus within 1.4 
cents of the union demands. With this information, the management could and 
did reach a settlement without coming to blows. It bargained the union down 
to the $.6 cents allowed under the standard. 

The examples of policy and planning decisions made with the aid of 
our business "clocks" by no means exhaust the possibilities. It is perhaps 
not too much to say that there is scarcely a single move in planning or 
policy which will not be substantially improved if it is made against the 
background of such standards as those shown in exhibit hi and exhibit iv. 


In the month-to-month control of operating costs and profit, the 
"clocks" have even more frequent use. They materially improve the ex- 
ecutive's capacity for "managing by exceptions" because they show the 
exceptions in specific dollar amounts, favorable or unfavorable. They 
make it possible to avoid situations like that in which a certain metal manu- 
facturing firm found itself. It had almost endless records of standard and 
actual costs by operations and by products. Yet the production manager 
had no single figure by which his over-all performance could be measured 
by his superiors. He explained a poor showing by saying, "Well, the vol- 
ume was down, you know." A good showing, naturally, was due solely to 
his managerial effectiveness! 

By contrast, when the standards for manufacturing are set up like those 
in exhibit v, the production manager as well as top management can see 
what ought to happen at any and every level of volume. Variable con- 
tingent costs are separated from rigid expense; and when compared to 
actual variable costs, managerial performance is measurable. No less im- 
portant is Column F in this exhibit, showing the standard manufacturing 
margin for each level of volume. 

exhibit vi and exhibit vii represent similar control sheets for selling and 
administration. For top management, exhibit viii is a simple one-sheet 
record that reveals monthly any "exceptions" or variance from standard, 
by major operating functions. To illustrate: 

From the right-hand column, top management sees at a glance that the profit 
for the year to date is, happily, $44,792 (Line 27) over budget (Line 25). 
What caused that "exception? " Who is responsible? And for how much? 

The exhibit itself breaks down the possible sources of the exception into: 
(1) external causes, i.e., production value under or over budget; and (2) in- 
ternal causes, i.e., expense under or over. I find it useful and even essential in 
discussion with associates to keep these two sources entirely separate and dis- 
tinct. Manufacturing and administrative managers feel no direct responsibility 
for a short-fall in sales; the sales manager feels no direct responsibility for 
variances in internal manufacturing and administrative expense. 

"Clocks" for Management Control 3 2 9 

Keeping this distinction in mind, let us examine the sources of the profit 
variance. These two items are the key: 

i. Line 22, gain over budget on realized production value $2i,55 2 

2. Line 23, reduction in internal expense 23» 2 4° 

Total profit variance $44J9 2 

The first, and larger, amount is traceable to external conditions. As Line 21 

discloses, the firm's production value was $63,611 over budget; at an operating 

margin of 33.88%, sales caused a profit gain of $21,552. 
The second source of profit gain is traceable to internal control. If we 

break down the single figure of $23,240 (Line 15), we can learn where the gain 

occurred, who is responsible, and for how much. Thus: 

Line 6, manufacturing expense $33» J 3^ 


Line 9, selling expense $7>54 8 

Line 12, administrative expense 2,348 9» 8 9 6 

Net gain $23,240 

Such a breakdown of the variances serves to fix the responsibility in concrete 
terms for each of the executives concerned. The problem of internal control of 
expense is not mixed and confused with the problem of overcoming external 

Whether the top executive needs or should carry the analysis still further is 
a matter of individual judgment and preference. For instance, should the sales 
manager be called upon to show the detail of his excess expense of $7,548, or 
should it be left to him to correct it in his own way? Either course may be 
desirable depending on the circumstances. Again, shall the top executive ask 
for a breakdown of the external variance in production value from budget 
(Line 21), say, by products, sales territories, or both? It is entirely possible, 
and in some instances desirable, thus to localize the market factors that led 
to this gain. 


How far should the senior executive go in the direction of finer and 
finer breakdowns? It is probably a matter of his individual judgment and 
of the capacities of his associates. In general, I find it preferable to avoid 
too much breakdown; at some not easily determined point it is likely to 
lift the expense of records maintenance beyond the gain to be had from 
the use of the information. Moreover, nothing paralyzes action on the part 
of executives like excessive analysis of figures. This procedure, like any- 
thing else, can be carried to extremes; there is a fascination about neatly 
tabulated figures and charts that needs to be resisted lest it lead managers 
to believe they are on top of their problems without thinking them 
through and coming to decisions. 

3^0 The Production Process and the Administrator 

A good general rule, in my experience, is to carry the analysis only far 
enough to give direction and stimulus to effective action on the part of 
the managers concerned. For that reason I find it best to begin with only 
the few and simple charts shown here; these usually are enough to enable 
the top management to maintain a perspective view of the business and 
to direct and encourage effective action on the part of managers down the 
line. In fact, in single-plant operations, these few charts will suffice per- 
haps indefinitely. In multiplant operations, similar records for each plant 
are desirable, with a composite set of controls at headquarters. 


What records are needed to maintain effective control? 
With the idea of limiting the number of records to those that can be, 
and will be, used profitably, I advise starting with two categories: 

i. Those records needed for business planning and major policy decisions. 
2. Those needed for month-to-month current control of costs and income. 

For management planning and easy, close watch on the long-term trend 
of business developments, the following data showing five years' figures 
at a glance may be invaluable: 

i. Breakeven production value. 

2. Safety factor. 

3. Date of reaching breakeven. 

4. Operating margin. 

5. Rigid expense. 

exhibit in shows these five master controls in tabular form as prepared 
for a metal products manufacturer. Of these five, the safety factor index 
is perhaps the central control which integrates the others. It is well to 
chart it along the lines of exhibit ii. Such a chart reveals at a glance any 
tendency to expand the rigid expense of added facilities and personnel 
faster than the realizable growth of the business, or conversely, any failure 
to keep up with the opportunities of the business. This might almost be 
termed the key to most other management planning decisions. A chart of 
the trend of the safety factor, along with at least a five-year table of all 
five master controls, as listed in exhibit hi by years, and perhaps by 
quarters for the current fiscal year, will usually suffice. 

For current month-to-month control of costs and income, my experi- 
ence is that the following limited number of records will adequately serve 
top-management needs: 

1. Four tables showing standards at all levels of production value, above and 
below breakeven: 

a. Standard total company pretax profit in dollars and expressed as a per- 
centage of production value and of net worth of capital used (exhibit i). 

"Clocks" for Management Control 3 3 1 

b. Standard total manufacturing costs (exhibit v). 

c. Standard total selling expense (exhibit vi). 

d. Standard total administration expense (exhibit vii). 

2. The following reports each month on standard comparative statement 
forms, showing the current month and year-to-date figures, which can be 
easily compared with records of preceding months: 

a. Key operating factors (exhibit in). 

b. Profit control, with a breakdown to show major operating functions, 
and reconciliation with budget to show the source of variances (exhibit 
viii ). 

These records usually turn out to be all that are really useful to top 
management in keeping a month-to-month perspective and having a basis 
for control. They will be enough to set up exhibit iv, which is the corner- 
stone of the control procedure advocated in this article (all the other 
controls and "clocks" are derived from it). Management will then be in a 
position to measure performance against common standards. Everyone 
will be able to have the same over-all perspective, and executive teamwork 
should soon achieve results of a high order. 


During the fiscal year, the total annual and the average monthly costs of 
certain items of expense change very little or not at all. Examples are real 
estate taxes, depreciation, property insurance, and executive salaries, 
among others. Such expenses accrue with the passage of time and do not 
fluctuate with changes in the volume of business done. They are rigid ex- 

In addition to such expenses, a plant also incurs costs for materials, sup- 
plies, power, labor, selling, and other items. The monthly and the annual 
amounts of such items will vary with the volume of business done, i.e., the 
amounts are contingent on volume or value of output. 

Our problem, therefore, is to separate the conventional accounts into 
the two categories of rigid expense and variable contingent costs. Our 
further problem is to state the total annual and the average monthly 
amounts of rigid expense in single dollar figures: then to state the variable 
costs in some definite mathematical relationship to the value or volume of 
business as, for instance, 66.12 % of production value. 


To solve the problem of separating rigid expenses and variable costs 
two approaches are commonly used. 

1. Simple Inspection. First we may inspect each of our expense accounts 
and decide which of them represent rigid expense and which represent 

ft* «< 



u-,vo r^- qs 

«"«•* rj w o 

cd so* <^> « 

O^ r^ "£ n" 

rj-00 «^ NO 


O (^ "4- on 
rh -tf- to so 
so «"*■> s r^- 

rfsO^CO & 
ON >-i rri u-i 
i-i r<-> rh tr-, 
rri rri rri in 

r^ r^ O "* 

m n rAfi 

i*-, Os r«"i l~~- 

^- M w OS 

rl r<~i r^i r<-i 



r^ •^■no Os <^j "j no ^ ff> 

m f^,oo <4 r^, o ^ 

N N «S 

msO m "ti ©> n 00 N toOO 
r^i^M ^-so GO CO On Os Os 

r* r-NO O n no vn rJ C> «^ 

roNO NO O <-<"> OsNO r<i OsNO 
r^. t£>o^ lo r^,Nq_ i-^ ^00^ <^ 

rj ^-M rr + io'riNM **■ 
w r-l so Os N "<*- t<-,NO NO r^ 

m rrii^NN wiflN CO NO 
CO Os r^,NO O m ro rt- Tf «o 
CO N ui O NO rr\ i-^GO^ to r|_ 

tF go" co~ r^ w^co" ■<£ i2 v 5f-^ 
n so Os o « w ^J^noco 
o\ m oo tn ww-j r^oo_ C£ O, 
p£ t? t£ u^no"so*~so"'no"~no'~ r^ 

rJ-O m Ni-ONO *N M 

m \o o no h-io o i^^n 

so i--~n in0>0 rrir ?^! u ? 
0^r^r^rT«^<^<^l^- 1 -' Wi 
3- m N rj-so r-CO NO «r> IT 

^f T? v^no^ r^ r^co co"co co 


























o o o 


o r^ «o o •flO 

m rt io r~~ 

O w 

M rr\ rr^ *T 


<S N 

M N N N 







SO r--CO Os O <- N r<-, Tf u~ 

"Clocks" for Management Control -> ->*> 

variable costs. Then these two categories may be listed in separate columns 
and the columns totaled. 

This is a simple method, but unfortunately the facts of business life are 
not so simply revealed, at least by the conventional arrangements of ac- 
counts on most operating statements. At best, only a handful of expense 
accounts are simon-pure rigid expense or strictly variable contingent costs. 
For instance, under the heading of manufacturing costs, raw materials, 
direct labor, and perhaps supplies may be strictly contingent costs— i.e., 
items rising and falling with the volume or value of output. Depreciation, 
of buildings and equipment, property insurance, and some others, may be 
strictly rigid. However, such expenses as power, supervisory and engineer- 
ing salaries, and maintenance will contain elements of both rigid and con- 
tingent costs. We cannot separate with sufficient precision the amounts of 
these two components in each account. 

Similarly, under the heading of administration and also under selling ex- 
pense, we find some strictly contingent accounts, such as cash discounts 
and sales commissions; we also find some strictly rigid expenses, such as 
warehouse rentals, bond interest, and cumulative preferred stock divi- 
dends. But many other accounts— sales salaries, sales travel, office clerical, 
and the like— contain elements of both rigid and contingent costs. The 
two components cannot be separated with accuracy by simple inspection. 

2. Standard Ratios. Some better method must be applied in order to 
separate precisely the great majority of accounts into their rigid and con- 
tingent components. This calls for constructing a table (or chart, or both) 
of standard total expense and standard total pretax profit for all levels of 
volume, such as that shown in exhibit i. The approach needed is mathe- 
matical. This is a formidable word and makes the process sound at first 
somewhat more intricate than it really is. Let me clarify the underlying 
principle by a homely and familiar example: 

In making coffee at home, many wives follow the old formula of one 
spoonful for the pot and one for each cup to be served; out of habit they 
allow two cups for each person. If at dinner four people come to the 
table, the total spoonfuls of coffee (t) required will be easily computed 
by this formula, where a is the constant one spoonful for the pot, and x is 
the number of guests: 

t = a + ix 
= 1 + 2.4 
= 9 spoonfuls 

The rigid expense in this familiar routine is, of course, "one or the pot" 
or a; the contingent or variable cost is the fluctuating volume in number 
of persons or x. 

Now, if the nominal head of the household has the temerity to do so, 
he may prepare a control standard showing the allowable coffee consump- 
tion for any meal, and also the average spoonfuls per person. He would 


The Production Process and the Administrator 

need only to construct a simple table in which the rigid a is replaced by i, 
and the variable x by the number of persons, with the factor 2 remaining 
constant, as follows: 

Number of 


Total spoonfuls 
(t = a + 2x) 

t = 1 + 2 • 2 = 5 

*= 1 + 2 • 3 = 7 

*= 1 + 2 • 4 = 9 

j = 1 + 2 ■ 5- = 11 

Average spoonfuls 
per person 



Note that: ( 1 ) total coffee used rises and falls with the number of per- 
sons (the volume) but not in direct proportion, because the rigid "one for 
the pot" is spread over an increasing (or decreasing) number of persons; 
(2) the average spoonfuls consumed per person diminishes with an in- 
crease in the number of persons, i.e., with a rise in volume, and increases 
with a decrease in volume. 

Executives will immediately recognize the parallel to production: the 
average total cost per order filled or per dollar of production value de- 
clines with an increase in volume and rises with a fall-off in volume. This 
is almost invariably true when total manufacturing cost, total selling 
cost, and total administration cost are considered. It is likewise true of 
many individual items comprising these or other categories of expense. 

Here, we may note, is the explanation of why the widely used com- 
parison of two periods on a basis of percentages to sales is inaccurate and 
misleading. The percentages will change with changes in sales as well as 
with the effectiveness of managerial control. And no one can say precisely 


(Ratio of actual production value to breakeven) 





•: : : : : : :: : : : :'; 




1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 


"Clocks" for Management Control 335 

how much of the change is caused by volume changes and how much by 
improvement or deterioration in managerial effectiveness, unless he is sup- 
plied with other data. 

For these reasons, the changes in amount of total manufacturing, total 
selling, and total administration expense cannot be anticipated solely by 
use of a uniform percentage applied to several possible levels of volume or 
output value. Nor can a pretax profit standard be fixed as a simple and uni- 
form percentage to sales at all levels. As sales increase, the percentage of 
total expense to sales will almost always decline (because the constant 
component of rigid expense— or "one for the pot"— is spread over an in- 
creasingly larger sales amount), and the percentage of profit will rise. 
Conversely, on declining sales the total expense percentage nearly always 
rises and the profit percentage shrinks. The same applies to profit on net 
worth (capital used), as exhibit i indicates. 

Hence, rigid expense must be separately determined with accuracy, so 
that the remaining variable part of any total cost can be stated as a constant 
percentage of production value at all levels within the range of normal 

"least-squares" equation 

Fortunately, the mathematicians have worked out equations that are, in 
the end, quite as simple as the one we have applied to coffee consumption 
at home. And they have also worked out the method for computing these 
equations from data available from a firm's regular books. The particular 
equation most often applicable is that called the "least-squares regression 
line." It takes the same form as our coffee equation: t = a -f- ex where t 
is total expense in any category, a is a constant representing dollar amount 
of rigid expense, c is a constant percentage, and x is volume or value 
which may fluctuate from month to month. 

In applying this method, it is necessary to arrange in tabular form, in 
separate columns, the monthly amount of sales or production value and 
the monthly amount of the expense of a given type, say administration, 
manufacturing, or a smaller subdivision. Some care may be needed to re- 
distribute over the months of a year expense items that often are charged 
out when bought or when paid: examples are real estate taxes and prop- 
erty insurance on a three-year premium basis. 

When the required tabulations are completed, the least-squares com- 
putations are then run out. 2 The end result is to separate total expenses 
into their two major categories, rigid expense and variable contingent 

2 The computational techniques are not given here. A statistics textbook should be 
consulted for the procedure. 


The Production Process and the Administrator 

Whether we want manufacturing costs, sales expense, administrative ex- 
pense, or total company costs, the least-squares equation expresses the 
rigid expense in a single dollar figure, and the variable contingent cost as 
a percentage of whatever production value (or sales) may be obtained. 
For instance, for the company in exhibit i the equation for figuring total 
expense (t) is: 

t — $1,247,332 + 66.12% of production value 

These items are what the top management needs to know. By substi- 
tuting in this equation any appropriate dollar amount of production value, 
multiplying that amount by 66.12%, and adding to the result the $1,247,- 
332 rigid expense, executives can obtain the standard total company ex- 
pense for particular volumes of business. Thus it is a simple matter to set 
up a table of standard total company expense for any level of volume. 

Measure of Gross Income. What measure of gross income should be 
chosen? In originating the formula for rigid expenses and variable costs, 
Knoeppel used net sales, and most other authorities have followed his ex- 
ample. However, sales usually means the invoice value of shipments. It 
does not mean either orders booked or goods produced. When there is 
any appreciable time interval between the booking of orders and/or their 
production and actual shipment, net sales is not by any means a suitable 
income measure against which to control variable costs. 

But there is an even more serious objection to the use of sales values, 
whether measured by orders booked, goods produced, or goods shipped. 
Sales value is definitely not a measure of any firm's internally disposable 
income. This is true because it consists actually of two sets of values: (1) 
the cost of raw materials, power, supplies, and so on, purchased from sup- 
pliers and not produced within the firm; and (2) the value of the firm's 
own productive effort in converting raw materials into its own finished 
product. Since money spent with suppliers cannot again be spent inside 
the business, a firm's internally disposable income is obviously limited to 
the value of its own productive effort, in short, to production value, which 
is truly 100% of the firm's economic contribution. 

Hence, production value is the ultimate limit of all internal business 
outlays, including corporate income taxes, dividends, and additions to 
capital account, exhibit iv illustrates this basic concept, beginning with 
Line 3. (Controllers will recognize here, as elsewhere in this article, the 
so-called "marginal income" type of accounting often applied to individual 
products or product lines.) I have yet to see a business in which manage- 
ment thinking and decisions are not substantially benefited by controlling 
internal costs and expenses strictly according to this one significant fac- 
tor—production value. 







6^ <s S^ u-, 



TC T o 



rf S, O <5 



w jj co 

<T ^.^ jr 

- S 




»*-> rt 11 ^ 

J* »M -C T\0~ 

OS 00 o Too 

4 -<3- *§ 
































■° bo 




00 o 

u , 

rt "13 







bo C 










<U o 


o o 
Ul -> 










a x 















C/3 CO 




00 00 




q no 


r^No oo 

m 00 

to oo 


O.NO uj 
ro unOo" 
00 NO On 


> NO 

E" - 

























o > 
8 hi 


° 6 -I 

O £ > 

<u rt d 

p «-i .O 

3-5 "§ 

"75 u P 


O x 

-B * - 

ai g M ^^ H w 

P 9 <u C S t/3 

-S'C > <u S c 

5 o OT3 X <u 

^ P - «h X Cu 

r-'OC P W X 

■8* 11.11 

.2 o SSc^< 

.ti-Q rt 
t« rt KrJ^^^v/-s 


c*a 8 

— ' > cu 
<u rt . 

C r< u 

P3 c rt 
rt bp« 







P *-" 

6 fi 

6 ul 

00 x 

bp a 

C rt 

•§ x 

8 8 

S3 a 

rt 4_> 

U i) 

u u 

vO O^ 

S^ o^ 

00 *ri 

00 r^ 

°°. NO 


m m 

ro fe^= 

r^ %^= 






d f 


rt <u 
P rt 

bo <u 

1) Ul 



The Production Process and the Administrator 


{Annual basis) 

Volume as 

percentage of 



(production _ 

value) Variable* 
(B) (C) 

Manufacturing costs 
allowable at standard 












$ 331,325 























100% (breakeven) 












































4, 1 43, 449 
































* Variable costs are 51.98% of production value in Column B; figures rounded, 
t Column B minus Column E. 


(Annual basis) 

Volume as percentage 

of breakeven 



(production value) 


Selling expense allowable at standard 





























100% (breakeven) 

































































* Variable costs are 5.66% of production value in Column B; figures rounded. 

"Clocks" for Management Control 



(Annual basis) 

Administrative expense 

Volume as percentage 

of breakeven 



(production value) 


allowable at standard 


Variable* Rizid 
(C) (D) 





$249,931 $98,544 




3, 1 2 9,M7 

265,553 98,544 




3,3 I 3,320 

281,173 98,544 





296,794 98,544 



100% (breakeven) 


312,414 98,544 





343,656 98,544 





374,896 98,544 





468,623 98,544 





546,726 98,544 





624,829 98,544 





676,519 98,544 





702,932 98,544 





718,553 98,544 





734,174 98,544 





749,795 98,544 


* Variable costs are 8.48% of production value in Column B; figures rounded. 


Current month 

Cumulative for year 

Production Value 



Sales Value of Output 
Less: Materials and Supplies 




Production Value 


$ 613,578 


to Budget 



Manufacturing Costs 

$ 409,196 



Variance from Standard 

$ 852 

$ 33,n6 


Selling Expenses 

$ 40,203 

$ 262,052 


Variance from Standard 

$ (1,922) 

$ (7,548) 


Administrative Expenses 

$ 60,243 

$ 392,679 


Variance from Standard 

$ (1,480) 

$ (2,348) 


The Production Process and the Administrator 

Total Costs and Expenses 

13. Standard $ 509,642 

14. Actual 512,192 

15. Variance from Standard $ (2,500) 

Pretax Profit 

16. Standard $ 103,936 

17. Actual 101,386 

18. Variance from Standard $ (2,550) 

Budget Comparisons 

19. Budget Production Value % 664,334 

20. Actual Production Value 613,578 

21. Variance from Budget $ (5°,75 6 ) 

22. Resulting Profit Variance $ (17,196) 

(33.88% of Line 21) 

23. Cost-Expense Variance ( 2,550) 

(Line 18) 

24. Total Profit Variance $ (i9,74 6 ) 





















25. Profit Required at Budget $ 121,132 

26. Actual Profit (Line 17) 101,386 

27. Total Profit Variance $ (19,746) 

It is to be preferred for practical as well as theoretical reasons. For in- 
stance, labor unions have come to demand an equitable "share of increas- 
ing productivity" as a part of their members' earnings. Production value 
supplies a true measure of "productivity" in the simplest and most useful 
manner I have yet discovered. Moreover, total payroll costs of hourly 
rated factory labor (direct and indirect, with overtime and premium pay- 
ments and "fringes") tend to bear a near-constant relationship to produc- 
tion value. Thus there is a basis for comparing this year's payroll with 
former years' despite changing levels of prices and productivity. 

To attempt to relate wages (or other variable contingent costs) to sales 
values invoiced or produced provides no sound basis for evaluating labor's 
demands. The practice rests upon the curious assumption that, somehow 
or other, material costs should maintain a constant percentage to sales 
values. Such can be the case in certain special situations; it is more often 
not the case, however, especially today when firms are changing from 
low-price to higher-price materials, and sometimes making subassemblies 
but at other times buying them. (Of course, far too many firms continue 
to price their products in such a way as to cause the material-cost com- 
ponent to whipsaw margins and pretax profits. Such pricing methods need 
considerable overhauling. But that is another story.) Besides, what is the 

"Clocks" for Management Control 341 

point of paying labor, salesmen, and others a commission on raw material 
costs necessarily included in sales values of output or shipments? 


It is now easy enough to set up five master control standards, as in 
exhibit in. For instance, when variable contingent costs are expected to 
absorb 66.12% of any amount of production value (taken as 100%), then 
only 33.88% remains as the operating margin. 

The operating margin is of overwhelming significance; it is certainly 
one of the five most important figures in any business. In this example, it is 
33.88% of production value, and is the only part of production value 
which can be used to pay rigid expense. In other words, we can make no 
profit until all rigid expense has been covered out of this 33.88% of pro- 
duction value. 

How much production value is needed to cover $1,247,332 of rigid ex- 
pense, and thus just break even? Obviously an amount which when multi- 
plied by 33.88% will yield exactly $1,247,332: 

$1,247,332 - 33.88% = $3,681,467 

But this is just a beginning: No one intends to be merely a "nonprofit" 
organization. If we are to make any profit this year, we must certainly do 
more volume than $3,681,467. Therefore, we come naturally to this con- 
clusion, and it is a highly important one: We must generate the break- 
even volume sowtetime before the end of the fiscal year, or there will be 
no profit. 

At what date before the end of the fiscal year should production value 
begin to exceed the breakeven point? This is the question which writers on 
this subject have neglected; in my experience it is a genuinely key point. 
Until it is answered, we cannot really have a true plan; we have only an 
objective but no time schedule for achieving it. Hence, we not only must 
set out to exceed the breakeven point, but must seek to do it at the earliest 
date which is reasonable in the light of current conditions. In my experi- 
ence, it is more vital to have all executives in an organization striving to 
shorten the number of months needed to break even than it is to have 
their eyes on some profit goal. For, once a firm has covered, in the first 
four, five, or six months of the fiscal year, its entire yearly rigid expense, 
profit follows almost as surely as water runs downhill. 

In exhibit in, then, the most probable date of passing the breakeven 
point is included (Line 3). This date serves to integrate time with volume, 
and to show the entire top management not only what is to be done but 
by what time. 

Let us go on to another significant use of the operating margin per- 
centage and of the breakeven point-forecasting the total production value 
needed to earn any stated amount of pretax profit. Suppose we want 

^ 42 The Production Process and the Administrator 

$1,000,000. I like to regard such a profit target as actually a "rigid ex- 
pense" of the business, since a part of it is, in effect, the cost of financing 
business growth and progress in the future. The remainder pays the own- 
ers for the use of their capital and for their income taxes. To obtain the 
amount of production value required for any amount of pretax profit, 
only two simple steps are needed: (1) add the pretax profit to rigid ex- 
pense and (2) divide the sum by the operating margin percentage. The 
result tells the executive, his associates, and his board of directors what 
it takes in volume to earn a normal return on net worth, to pay regu- 
lar dividends, to pay "extras," and to provide any other funds which 
were included in the $1,000,000. Thus the volume in production value 
needed is: 

$1,247,332 + $1,000,000 = $6,633,211 

We can also compute the percentage of our actual volume to break- 
even volume and find the safety factor (Line 2, exhibit hi). This factor 
tells us how much volume we can lose before going "into the red." In 
fact, almost anything an executive wants to know about the operating 
economics of his business can be learned readily by the use of breakeven 
point controls. 



Frank J. Andress* 

As a new product continues to be produced in a consistent fashion the 
improvement in labor efficiency can be predicted. The "learning curve" 
device has been developed to aid in the intelligent use of this in planning. 

Practically every manufacturing company has to forecast labor time 
and cost per unit of product in order to set selling prices, plan delivery 
schedules, calculate capital and labor needs, and so on. For years fore- 
casting labor input has been a laborious and time-consuming job. Now, 
with the aid of a new concept developed by the aircraft industry in 
World War II— the learning curve— such forecasting should become much 
easier, quicker, and more accurate. 

* From Harvard Business Review, XXXII, 1 (1954), 87-97. Reprinted by permission 
of the Harvard Business Review. 

Learning Curve as a Production Tool 343 

That the learning curve concept can be applied readily to common 
types of manufacturing operations outside aircraft is illustrated dra- 
matically by the following incident: 

During World War II an executive of a home-appliance manufacturing 
company chanced to cross paths with an executive of a large West Coast air- 
craft firm. The appliance executive mentioned that it had taken his company 
two years to determine the exact cost of the electric refrigerator which it 

The aircraft executive pointed out that in many cases his company had been 
forced to determine costs on similar items in a matter of a few minutes, and 
said, "I'll bet you a steak dinner that I can predict the cost of your 100,000th 
refrigerator within 10% accuracy by using a learning curve based on aircraft 

The manufacturing executive accepted the bet. The only information he 
furnished was the weight of the refrigerator and the cost of the first unit pro- 
duced. During the next few minutes he watched while the aircraft executive 
worked with pencil, ruler and log-log graph paper. 

When he had completed plotting the curve, the aircraft executive stated: 
"Your 100,000th unit should cost you $162.50." 

"Just drop the 50 cents," the appliance executive said. "It was actually 

In the following pages, I should like to explain how the learning curve 
works, point out some of the problems that it poses, and discuss its po- 
tential applications in different industries. 


The basic theory of the learning curve is simple: a worker learns as he 
works; and the more often he repeats an operation, the more efficient he 
becomes, with the result that the direct labor input per unit declines. 
This holds true whether the industry is aircraft, metalworking, textile, or 
candy-making. What was not known until a decade ago is that the rate 
of improvement is regular enough to be predictable. It is this fact that 
makes what would otherwise be a rather commonplace observation the 
clue to a broader and more practicable concept for business. 


Even before the Second World War, when the theory came into gen- 
eral use in the aircraft industry, it had been recognized that the direct 
labor input per airplane declined with considerable regularity as the 
cumulative number of planes produced went up. (Because of the high 
unit costs of a fuselage or complete plane, a phenomenon of this sort 
would naturally be more conspicuous in the case of aircraft than of most 
other industries.) Not only did this mean that the unit cost progressively 
declined; but of more importance, particularly in wartime, it meant that 
more planes could be produced with the same work force and facilities. 

344 The Production Process and the Administrator 

This reduction in labor time per unit might have been called rising 
productivity. On the other hand, the process repeated itself whenever a 
new type of plane was put into production; that is, the direct labor input 
reverted back to approximately what it had been when the first plane of 
the preceding type was put into production (assuming the two plane types 
were of similar size and configuration). The phenomenon was referred to 
as learning because of this repetitive characteristic, rather than as produc- 
tivity which implies some sort of sustained improvement. 

Regardless of the label used, the significant fact is that the learning pat- 
tern occurred with considerable regularity. It was not long before the 
various airframe companies had established certain standards of learning 
which they used as a basis for predicting direct labor input. Then the 
armed services became interested and sponsored a statistical study of direct 
labor input by the Stanford Research Institute covering a majority of all 
aircraft produced in the Second World War. (The industry lent itself 
well to such a study because most aircraft companies operate on a job- 
shop basis of cost accounting, which lessens the task of obtaining the nec- 
essary data.) 

As a result, there was developed a series of learning curves which repre- 
sented the average experience for various categories of airframes— fighters, 
bombers, and so forth. Although these curves were all different in terms of 
their starting points (i.e., the labor input for the first plane of a particular 
type), the great majority had one characteristic in common: their rate of 
improvement. It was this fact, essentially, that started speculation about a 
general theory of learning curves. 

The rate of improvement which was found to hold true for the opera- 
tions covered by the survey was such that, once production on a plane 
got going, the 4th unit required about 80% as much direct labor as the 
2nd; the 10th, 80% as much as the 5th; the 200th, 80% as much as the 
100th; and so forth— in each case a reduction of 20% between doubled 
quantities. Because this rate of improvement seemed to prevail so con- 
sistently, it was concluded that the aircraft industry's rate of learning was 
approximately 80% between doubled quantities. That standard is applied 
to this day in analyzing a variety of procurement, production, and costing 
problems within the industry and within particular companies. 


So that the reader may obtain a better picture of the learning curve, a 
hypothetical one is reproduced in exhibit i. Here we see two lines drawn 
on log-log graph paper such that the 20% reduction holds true between 
all doubled quantities (except for a small distortion at the upper end). 
The lower line represents the unit curve (the direct labor hours for a par- 
ticular unit) while the upper line represents the cumulative average 

Learning Curve as a Production Tool 345 



12 5 10 20 50 100 200 500 1,000 


curve (the average direct labor hours for all units produced up to any 
particular point). Thus the unit curve indicates that the iooth airframe 
should require approximately 15,600 hours, whereas the average for the 
first 100 airframes produced should be approximately 23,000 hours. These 
two lines can, of course, be expressed mathematically. There are three 
learning curve formulas: 

(1) Y = KX n , where: 

Y = cumulative average man-hours for any number of units, 

K = number of man-hours to build first unit, 

X = any number of completed units, and 

n = log (% of learning curve) -f- log (2). 

[Note that the exponent n will always be a negative number unless the 

slope is 100%. Such a negative exponent can be handled as follows: 

x~ n = 1 -j- x n .~] 

(2) U = (n + \)KX\ where: 

U = unit man-hours for a specific unit, and all other symbols are the same 
as above. 

[(n + 1) is referred to as the conversion factor, and for an 80% curve: 
(n+ 1) = + 0.67807.] 

(3) T = KX n (X), where: 

T — total man hours required to build a predetermined number of units, 
and all other symbols are the same as above. 

The cumulative curve in exhibit i appears as a straight line because of 
the nature of log-log graph paper. (So does the unit curve after its initial 
jog.) A learning curve is usually drawn on log-log paper simply because 
in straight-line form it is much easier to project. 

34 6 

The Production Process and the Administrator 


5 10 15 20 25 30 35 40 45 50 


If the same curve were drawn on ordinary graph paper it would become 
a true "curve," as exhibit ii indicates. This curve, of course, more dra- 
matically illustrates the absolute amounts of reduction from one unit to 
the next. Thus, though the percentage of reduction remains the same, it 
applies to a progressively diminishing base, and the absolute amounts be- 
come less and less until they virtually level off. This reflects the fact that 
in actual experience the process of learning a given operation eventually 
approaches a plateau where relatively little further improvement takes 


What factors account for the learning curve phenomenon? A distinc- 
tion must be made between (a) learning in the literal sense, on the part 
of both workers and management, but primarily the former, and (b) a 
whole series of other factors, among which management innovations ap- 
pear most significant. These causal factors operate sometimes in combina- 
tion and at other times in opposition. To this extent, the learning curve is 
more of an empirical method for charting all the various forces which 
work on labor hour input than it is a truly scientific device. 

The significant fact is the consistent behavior of the curve, which in- 
dicates that of the various factors learning in the literal sense is the pre- 
dominant influence. It results in a smooth pattern of labor hour reduction, 
whereas the other factors can be erratic. Moreover, the more opportuni- 
ties there are for learning, the steeper the curve, as the experience of the 
aircraft industry demonstrates: 

Learning Curve as a Production Tool 


Approximately 75% of the total direct labor input in the industry is as- 
sembly; the balance is represented by machine work. In assembly work there is 
a relatively large scope for learning; in machine work the ability to reduce 
labor hours is greatly restricted by the fact that the machines cannot "learn" 
to run any faster. Accordingly, when the proportion of assembly work is less, 
the reduction of labor input is slower. For example, in the case of operations 
made up of approximately three-quarters machine time and one-quarter as- 
sembly time (the reverse of the usual situation in the aircraft industry), the 
approximate rate of learning has been found to be 90% rather than 80%. That 
is, the labor hours drop only 10% between doubled quantities, compared 
with 20% for the industry generally. 

Of course the nonlearning factors do have some influence. That is why 
learning curves in reality are not as smooth as the hypothetical one in 
exhibit 1. For example, new machinery introduced by the company will 
bring about savings in labor hours; so may time studies or design changes. 
The result will be to give a jerky effect to the curve, as exhibit hi illus- 

The irregular line is the actual learning curve experienced in the pro- 
duction of a certain type of fighter plane, whereas the straight line is the 
industry average for all fighters established by the Stanford Research 
study (in this instance expressed in hours per pound of airframe). The 


100 1- 

50 - 

1.0 - 



2 5 10 20 50 100 200 500 1,000 


348 The Production Process and the Administrator 

actual curve is quite smooth until the 250th plane; it has a slope of ap- 
proximately 77V2 %, which corresponds closely with 78% %, the industry 
average for this particular type of plane. After the 250th plane, however, 
it drops off sharply. The implication is that learning in the literal sense 
was predominant in the beginning (or else a whole variety of other factors 
canceled each other out very neatly) ; but subsequently other factors en- 
tered the picture, and the curve became quite unpredictable. 


There is no sure-fire accuracy about the learning curve. If management 
is to get accurate forecasts with it, certain pitfalls and limitations must be 


Perhaps the greatest need for caution is in the selection of labor hour 
data to be used in plotting the curve. For instance: 

1. How is increased utilization of purchased parts to be handled? If a 
manufacturer receives a larger portion of his raw material in a finished state, 
his labor input per unit should decline. However, the decline is merely the 
result of shifting labor input from his plant to that of his supplier, and it 
would be erroneous to show the decline as a real reduction in the labor input 
per unit. The same work is being performed in total, and no net saving nas 
been realized. 

2. Similarly, it might be possible to generate a direct labor reduction of 
1,000 hours by spending 10,000 hours on additional tooling and engineering. 
Of course there would be no real saving as a result. Clearly, therefore, direct 
labor hours cannot be considered as separate from the changes in the other 
elements of cost. 

There is also the danger of labor "savings" resulting from a mere re- 
shuffling of the accounting records. The learning curve plots direct labor 
only. But what happens when greater use is made pf supervisory labor, 
which is classified as indirect? If the emphasis is placed on direct labor 
savings alone without considering overhead and, particularly, indirect 
labor, a very distorted picture of direct labor productivity is likely to, r§= 

Another problem to watch out for is that of a change in the labor 
"mix." By hiring more skilled and correspondingly more expensive work- 
ers, a manufacturer may bring about a reduction in direct labor hours only 
to face the irony of an increase in direct labor cost. 

Finally, changes in direct labor hours may result from changes in the 
rate of production. If the rate of production is increased from below 
"normal" levels to, say, 90% of capacity, it may be possible to organize 
the work force in a more efficient manner and thereby reduce the labor 

Learning Curve as a Production Tool 349 

input per unit. In other words, there is the possibility that a decline in the 
direct labor input per unit may be due almost entirely to increased vol- 
ume, not to "learning." 

The point of all this is that the computation of learning curves is sub- 
ject to many hazards. There are technical methods for taking these hazards 
into account and compensating for them; there is no great difficulty on 
that score. But from the management viewpoint there is a real problem 
in the constant need to be on guard against errors creeping into the data 
for the learning curve and distorting it. The responsibility for this rests 
with the executives in the control function, collaborating with account- 
ants, plant superintendents, and industrial engineers. 


That these hazards were avoided in the study of the aircraft industry's 
experience— in other words, that the 20% reduction in labor hours be- 
tween doubled quantities was not merely the result of inaccurate compila- 
tion of data or of a failure to isolate accurately the factors which actually 
produced the reduction— is indicated by the following points: 

1. The direct labor input data measured consisted not only of direct labor 
performed by the prime contractors, defined as on-site, but of labor performed 
by subcontractors as well, defined as off-site. In other words, insofar as possible 
the total labor input was measured rather than merely the prime contractor's 
labor input, thereby avoiding any apparent savings resulting from the in- 
creased utilization of purchased parts by the prime contractors. 

To be sure, the measurement of total labor input is practically an impossi- 
ble task when routine standard parts must be accounted for. Thus, a sub- 
contractor using 1,000 ordinary washers may have manufactured them or 
procured them from a local mill supply house. Nevertheless, it is fair to 
assume that over short periods of time, such as six months, the type and the 
quantity of standard parts used (e.g., the washers) would not vary greatly; 
so that, whether the practice were to buy or manufacture them, the figures 
used to compute the learning curve would not be affected first one way, 
then the other, during the period. 

2. We know that time studies were used to check similar operations— 
especially the important assembly operations— at different periods in the 
learning cycle. These time studies indicated a pattern similar to that shown 
in EXHIBIT 11. 

3. Purchasing agents, production managers, methods engineers, and ac- 
countants in the aircraft industry have applied the learning curve principle 
for ten years now. Presumably these men would have discovered the "joker," 
if there were one> in this period of time. Significantly, its use has steadily in- 
creased in recent years. 

4. It is difficult to account for direct labor hour savings because of other 
factors— for example, excessive expenditures on tooling and engineering. Yet 
in a proflt-and-loss economy it is hardly likely, for instance, that a manufac- 
turer Would intentionally spend $500,000 on excessive tooling in order to effect 
a $100,000 reduction in direct labor. 

->e The Production Process and the Administrator 

All factors considered, the learning curve seems to have stood the test 
of time in the aircraft industry as a valid, reasonably accurate method of 
forecasting direct labor requirements. The fact remains that it cannot 
properly be called— at least in its present state— a scientific tool. It may 
one day become that, however, and in the meantime there is the prospect 
that it can continue to serve management quite usefully in its present form. 


Some of the specific ways in which management in a wide variety of 
industries can use the learning curve— from getting new help on "make or 
buy" decisions to planning closer delivery schedules— are apparent from 
the experience of companies in and associated with the aircraft industry. 
A discussion of these uses will doubtless suggest others to readers. 


The learning curve is continually used by the Air Force, Navy, and 
aircraft companies to forecast direct labor hours for a predetermined 
number of airframes as a basis for negotiating prices. Here are two illustra- 
tive cases (the real names of the companies cannot be used) : 

The Standard Aircraft Company. This company had just finished construct- 
ing 221 airframes under a postwar contract with the Air Force. Its actual 
learning curve, authenticated by an audit, was as shown by the solid line in 
exhibit iv. The 221st ship required about 17,750 hours (2.42 hours per pound, 
the figure indicated by the curve, times 7,334, the number of pounds per air- 

At this point the Air Force wanted to purchase another 333 airframes of 
exactly the same type. The question was not whether to use a learning curve, 
but rather which one of two would be better— the original curve based on 
World War II experience that had been used to price the original contract 
(see the broken line) or a new one based on a continuation of the figures for 
the 221st plane. Because the World War II curve would have produced an 
average of 17,730 hours for the 333 additional planes— obviously too high since 
the actual record had already reached almost the same figure for the 221st 
plane— the decision was made to drop it and instead to project a new 80% 
curve from the 221st plane to the 554th on the assumption that the rate of 
learning would continue during the production of the additional 333 planes. 

On this basis, an average of 14,660 hours was obtained for the 333 planes. 
The figure finally agreed upon between the two parties was 14,450 hours (a 
small change in design being computed separately). 

The Webster Machine Company. Following the outbreak of the Korean 
War the Army's own facilities were inadequate to meet the demand for gun 
barrels. Accordingly, the work was "farmed out" to private companies, one of 
which was the Webster Machine Company. 

The contract called for a single rough boring operation on 2,000 gun barrels; 
in other words, machine time (rather than assembly time) represented a large 

Learning Curve as a Production Tool 






i2.0 h 





O 0.5 







10 20 50 100 200 500 1,000 


part of the work. The initial contract price was $76.70 per unit subject to price 
redetermination. When the price was redetermined, Webster had completed 
545 units at an average direct labor cost, verified by audit and time study, of 
35.3 hours per unit. What direct labor cost should be charged in computing the 
price for the uncompleted units? Both parties recognized the fact that the 
3 5. 3 -hour figure was too high for the contract as a whole and that because of 
some sort of learning factor or increase in the contractor's "know-how," the 
new figure should be lower. 

Webster argued for 22 hours per unit for the uncompleted units, pointing 
out that the Army had seven years' experience in this operation whereas 
Webster had never done it before, exhibit v shows the company's proposal in 
the form of a learning curve. The figure of 35.3 hours was recognized as the 
cumulative average for the first 545 units and, accordingly, was plotted as one 
point on the graph. To get the second point, which was necessary in order to 
construct the curve, the company computed the cumulative average for all 
2,000 units (including the 545 produced to date) on the basis of 22 hours per 
unit for the uncompleted units; and thus arrived at an over-all average of 
25.62 hours. A straight line drawn between the points had a slope, or rate of 
learning, of 85% between doubled quantities. Applying the conversion factor 
for an 85% curve, which is 0.766, to the 25.62 average for all 2,000 units, it ap- 
peared that the direct labor theoretically required for the last unit, according 
to Webster's proposal, was 19.6 hours (25.62 X 0.766 = 19.6 hours). 

The Army supply officer took a very different view. He argued that since 
the Army was performing the operation itself in 11.6 hours, 13 hours was a 
fair figure for the uncompleted units. (Webster did not possess the Army's 

35 ; 

The Production Process and the Administrator 

special hoists.) His proposal, which was computed in the same manner as 
Webster's, is also shown in exhibit v. The slope of his line was a phenomenal 
71.5%. In order to obtain the cumulative average of 19.08 hours for all 2,000 
units called for by his proposal, Webster would have had to produce the last 
unit in 9.82 hours (19.08 hours X 0.515, the conversion factor for a 71.5% 
slope). This was 1.78 hours less than the time the Army was taking to produce 
the gun barrels, and the Army had seven years' more experience on the job. 
Clearly, Webster's proposal was the more reasonable, especially in view of 
the fact that machine time represented the greater part of the work which, 
judging from experience elsewhere, meant a slower rate of learning (or a higher 
percentage of labor time required between doubled quantities). The parties 
finally agreed on an 85% curve. 



Q£ 10 




1 400 1,000 2,000 10,000 




545 units @ 35.3 hours = 19,238 hours 
1,455 units @ 22 hours = 32,010 hours 

51,248 hours 
51,248 -T- 2,000 = 25.62 hours 

Learning Curve as a Production Tool 353 


545 units @ 35.3 hours = 19,238 hours 
1,455 units @ 13 hours = 18,915 hours 

38,153 hours 
38,153 — 2,000 = 19.08 hours 

Of course in many applications the usefulness of the learning curve for 
pricing decisions can be extended even beyond the important element of 
direct labor costs. To the extent that labor charges form the basis for 
allocating other costs, the curve also will serve as a guide to determining 
factory overhead, general and administrative expense, and so on. 


Make or buy problems are common to practically all manufacturing in- 
dustries. The decisive issue is usually an economic one— whether it is 
cheaper for the company to make an item itself or procure it from the 
outside. The learning curve can be a helpful device for resolving such an 
issue, particularly when buying means subcontracting or ordering on a 
negotiated price basis. To illustrate (again with company names dis- 

Early in 1952 the Lee Aircraft Company was faced with a cutback in its 
production as a result of the Air Force stretch-out program. Consequently, it 
was inclined to cancel some of its subcontracts and pull the work back into its 
own shop to keep it fully occupied. 

One subcontract which it thought of canceling was with the Roberts Manu- 
facturing Company for 372 landing flap assemblies— an item which it also was 
manufacturing in its own plant. To arrive at a comparison of its own and 
Roberts' costs of manufacturing the assemblies, Lee decided to plot the re- 
spective learning curves. 

Lee had already produced 165 assemblies, with a figure of 445 hours for the 
165th unit, and was well along the downward slope of its learning curve; con- 
tinuation of the curve indicated a total labor input of 111,000 hours for 372 
additional units. In comparison, the Roberts Company, while apparently a more 
efficient producer of the item, was just getting started on its learning curve; if 
it went on, it would be able to produce the 165th unit at an expenditure of 402 
hours— 43 hours less than Lee— but continuation of its curve from the earlier, 
higher point at which it then was indicated a total labor input of 164,000 hours 
for the 872 units, or 53,000 more than Lee. 

The foregoing analysis served to pinpoint the question for management's 
judgment. In the short run, it was more economical for Lee to cancel the 
subcontract and do the work itself. In the long run, however, it would be less 
expensive to leave the work with Roberts inasmuch as it could produce the 
landing flaps for about 10% less labor since it had got as far out on its learning 
curve as Lee. Therefore, the decision hinged largely on the probable total 
future demand for landing flaps of this type. Since this total future demand 

1*4 The Production Process and the Administrator 

was difficult to measure, Lee decided to take advantage of the direct labor 
savings offered at the time, which amounted to over $300,000. Accordingly, it 
canceled the contract with Roberts. 


The time required to perform an operation or a whole series of opera- 
tions is a vitally important factor in management decisions about produc- 
tion equipment, number of workers, work flow, production control, and 
practically every other aspect of production. Since the learning curve is a 
means of forecasting the time required, it can often be a helpful tool. 

Perhaps its simplest and most obvious use is in forecasting output. For 
example, if the size of the work force is kept constant, the declining trend 
of labor hours resulting from learning will result (assuming that materials 
and purchased items are available on schedule) in a rising trend of deliv- 
eries. If, say, in March the average direct labor per plane is 20,000 hours, 
and by May it is down to 15,000 hours, then a work force of 2,000 people 
working a total of 400,000 hours a month can supposedly produce 20 
planes in March and 26.7 planes in May. If changes in the work force are 
anticipated, they can of course be taken into consideration and the fore- 
cast altered accordingly, 

A more complicated but no less important use of the learning curve is 
in situations where sales are fluctuating and management needs to plan 
the proper size of the work force in advance. Here is a hypothetical ex- 

A manufacturing company determines that the demand for its new self- 
propelled diesel railroad car calls for a production schedule like that in the 
three lef thand columns of exhibit vi, after allowing time for tooling, engineering, 
and procurement. Its first step would be to work out a learning curve for 
these units, with the production schedule representing the point of departure 
and the size of the work force the dependent variable. 

If past experience indicated a rate of learning of 80% between doubled 
quantities, as in the aircraft industry, it might then arrive at a learning curve 
something like the one shown in exhibit vii. The monthly delivery schedule is 
indicated by the vertical lines; the points where they intersect the "curve" rep- 
resent the cumulative averages of direct labor hour requirements as of the end 
of each month. 

By taking these averages from exhibit vii and multiplying them by the 
cumulative number of units produced, the cumulative total hours for all pre- 
ceding months can be obtained (see the fourth and fifth columns, exhibit vi). 
The differences from month to month represent the total direct labor hours 
for any one month (sixth column, exhibit vi). 

If the latter figures are divided by the number of hours worked per month 
(in this case 200), the needed number of direct employees can be obtained. 
Thus, it can be seen from exhibit vi (righthand column) that about 2,228 em- 
ployees are required in the first month, and about double that number in the 
last month, whereas the number of units produced per month increases almost 
nine times. 

§ 8 O 

•si © 

^ i 

QOM iriffi u-iOO l«-i in 
N i-O >n tJ- r^ m tJ- 

is rT H pT m en r?i 'f 




TsO> h o o* m r-^ o* 
•3- Tf 4- t,*o no r^oo 



o o o o o o 
o o o o o o 
•^ O r O r ^ ^ o, 
in insO^^O in "00 

tJ- r-~. >-c oo >-i i-i t^v 

■3-00 mOO m 



*C °, ^ "CC ^ r "C 

O* mOO Tt - <-t O* t*» in 


m ^- r^ m so i»im 

m m n rn 

n Mtf m 1 *© r-oo 


The Production Process and the Administrator 



1 2 5 10 20 50 100 300 


Finally, the figures on total direct employees can be broken down into the 
various jobs such as fabrication, subassembly, and so forth. Knowing from the 
master schedule when each of these types of work must start, management can 
estimate the necessary training time and, working backward, draw up a hiring 
schedule by types and quantities of workers. 


To the financial man responsible for raising the cash necessary to finance 
operations, the learning curve is a help because it affords a basis for com- 
paring prices and costs and thus for estimating the period of financial 
drain when expenditures exceed receipts. Such information enables him to 
take any necessary steps in advance for special arrangements with custom- 
ers or with the bank. One way that appropriate action can be taken in 
advance as a result of a learning curve forecast is illustrated by the fol- 
lowing case (disguised): 

The Emmons Company had encountered considerable trouble because its 
suppliers found that they lost money in the initial phases of their production. 
In one instance involving a contract for 300 units, the supplier found himself 
"in the hole" for the first 90 or 100 units since his costs exceeded the selling 
price. The Emmons purchasing agent recognized that this was quite natural 
and explained it by means of the diagram in exhibit viii. (It is assumed, for 
illustrative purposes, that total cost and direct labor cost are synonymous.) 

Learning Curve as a Production Tool 






10 100 



The purchasing agent suggested that, if the supplier could not sustain the 
drain during the production of the first 90 units, the contract be broken 
down into two or more contracts with successively lower unit prices based 
upon the successively lower average costs (see exhibit ix), thus helping the 
supplier over the initial high costs of production. 

This problem could also have been handled by one contract containing 
progress payments (cash payments based upon the contractor's incurrence of 
costs), instead of using two separate contracts. Under either method, however, 
Emmons would have had to use some device such as the learning curve to size 
up the financial implications of changing production costs. 






| 50 










^^^^ ,UNIT COST 

^^^^""""^ CURVE 






1 1 L 



1 "<^ 




10 50 100 




The Production Process and the Administrator 

The pricing, production, financial, and policy applications of the learn- 
ing curve discussed in this section do not by any means cover the field. 
There are many other applications. For example, wage incentive plans 
could be established for groups of workers based on the steady improve- 
ment resulting from "learning"; or procurement schedules might be syn- 
chronized more closely with production needs, with resulting reductions 
in warehouse charges. 

Rather than attempt to suggest the wide variety of specific applications 
in more detail, however, it might be better at this point to discuss some of 
the underlying criteria which individual companies could use in making 
a definitive appraisal of the practical possibilities of the learning curve 
for their own situations. 


There is every indication that the learning curve offers a practicable 
answer to the needs of thousands of manufacturing companies for fairly 
accurate forecasts of direct labor requirements and productivity, but it is 
still a new device in a more or less "experimental" stage. The forward- 
looking management that is interested in the learning curve today needs to 
know essentially two things: (i) Is it applicable to the particular situation 
of the company? (2) If so, what steps should be taken to construct prac^ 
ticable learning curves? 


Judging from what we know to date about the learning curve, its use- 
fulness depends on these factors: 

1. Product innovation- -In companies where major and minor design changes 
are frequent, where new products are often introduced, or where manufactur- 
ing is characterized by short runs at well-separated intervals, "learning" is an 
important factor in the performance of the operators. Such companies are near 
the "top" of their learning curves, where savings are significant between units 
of production, a great deal of the time. Accordingly, accurate forecasts of 
direct labor time are especially helpful. 

2. Proportion of assembly time— As we have seen, the more an operation is 
made up of machine time, as opposed to assembly time, the slower the reduc- 
tion of labor time is likely to be, since the rate of learning is necessarily a 
smaller part of the picture. Thus, the aircraft industry has found that there 
is about a 20% or greater reduction in direct labor time between doubled 
quantities in the assembly of a wing, but only 10% in certain kinds of metal- 
boring operations. 

Learning Curve as a Production Tool 359 

3. Advance planning— The, more the pressures of immediate production 
can be overcome so that an operation can be planned in advance, particu- 
larly in the way of methods analysis and tooling, the more predictable the 
rate of reduction in labor time will be. "Learning" in the literal sense tends 
to produce a smooth curve; by contrast, changes in methods, tooling, and 
so forth during a production run will make the learning curve "jerky" and, 
at the same time, probably give it a more pronounced slope. 

On the basis of these criteria, industries of all sorts, in addition to air- 
craft, ought to be able to apply the learning curve profitably. Here are 
just a few possibilities: 

1. Electronics—Here product innovation is extremely important, and the 
long, complicated assembly lines attest to the importance of assembly labor. 
The same conditions exist that have made learning curves helpful in the 
aircraft industry. 

2. Home appliances- The moral of the story in the introduction to this 
article about the refrigerator manufacturer and the aircraft executive ought 
to apply with equal force to an endless variety of other products, from air 
conditioners to dishwasher-sink combinations. 

3. Residential home construction— A large-scale builder of residential prop- 
erty should experience a decline in the unit cost of a house with successive 
units completed. The direct labor input on houses constructed on the site from 
basic materials represents, of course, a substantial percentage of the total cost. 
Moreover, large-scale contractors produce many basically similar units; and 
they use semiskilled, noncraftsman labor to a large extent, resulting in fewer 
union controls over output. Accordingly, they should find the learning curve 
useful in making up bids and forecasting the number of workers required, 
capital needs, and so forth. 

By contrast, the learning curve would be of dubious benefit to a small con- 
tractor who builds but one or two similar units and has, therefore, very little 
basis for making a projection. The small contractor is often inhibited, further- 
more, by union limitations on the output of skilled craftsmen. 

4. Shipbuilding- -In this industry a large percentage of the total cost is- repre- 
sented by direct labor. Moreover, experience in the construction of basically 
similar bulkheads and panels provides information on direct labor costs which 
can be projected to cover the units contemplated for production. Note that 
different sizes and variations of basically similar items can be compared for 
learning-curve purposes by using the relative weights of the units as a basis 
of comparison. 

5. Machine shops— Even in plants where a large proportion of the work is 
machining rather than assembly, the learning curve promises to be helpful. 
Earlier in this article mention was made of a case in which a learning curve 
was used in negotiating the price of gun barrels manufactured by a machine 
company. Aircraft companies have also applied the learning curve profitably 
to the production of subcontractors making such products as finished castings. 

On the other hand, it seems apparent that some industries would find the 
learning curve of little value. In the case of basic chemicals, plastics, and 
petroleum refining, for instance, the direct labor element is negligible; in 
other cases— like the manufacture of some kinds of standard toys— there 
is a high ratio of assembly work but very little product innovation. 



The Production Process and the Administrator 

If the learning curve principle looks promising to management in view 
of its particular circumstances, the next problem is to construct one for 
actual use. Here is the approach to take in doing so: 

1. The first step is to obtain plottable data on number of units produced, 
which is fairly easy, and on direct labor hours, which may be quite difficult. 
Where the various hazards mentioned earlier prevail, the reduction of labor 
hours may be either illusory or fortuitous. Moreover, because actual direct 
labor hours are extremely difficult to obtain from a standard cost system, it 
will often be found necessary to use periodic time studies or a special job-cost 
system of accounting. 

2. Once the necessary company data have been collected in an accurate 
form, they should be plotted in order to determine what the approximate 
slopes are for the different operations. Slopes on subassemblies are particularly 
desirable since they can be used in the construction of learning curves for new 
end products using some of the same or similar parts as products previously 
produced. In any event, the slope is important since starting points for the 
curves will usually vary from one design to the next whereas the slope or rate 
of learning will, judging by past experience, remain relatively constant. 

3. Finally, it should be observed that the nature of the learning process is 
such that when the curve is plotted on log-log graph paper, it should approxi- 
mate a straight line. For experience indicates that the straighter it is, the closer 
it is to depicting the true nature of the learning process. 


If one were to attempt to look back at the past and present from some 
future date and put the development of the learning curve in perspective, 
he might see it as a three-stage process in which the present is but the 
middle stage. 

In the first stage, learning was recognized by management as a fact of 
industrial life, but allowances were made for it (if at all) only by hunch 
or by "guesstimate." No doubt some of these allowances came pretty 
close to the mark. In a textile mill, for example, the production manager 
allowed a 13 -week period for learning by the work force on a certain 
military contract; in figuring the hiring schedule required to meet the 
delivery schedule, he estimated that the workers, who were on the Be- 
deaux system of incentive wages, would improve their production from 
40 "B's" (Bedeaux performance units) per hour to 80 "B's." By astute ob- 
servation he was able to approximate the forecasts that a learning curve 
would have given him. 

In the second stage learning curves were plotted for the first time (in 
the aircraft industry) and specific rates of learning, usually in the neigh- 
borhood of 80% between doubled quantities, were discovered to exist 
for many operations. Nevertheless, valid generalizations have remained 

Learning Curve as a Production Tool 361 

suspect, especially about the behavior of curves in other, untried cir- 

In the final stage, the future historian might conclude, learning curves 
were tested on a wide scale, refined, and many things discovered about 
them that were hitherto unknown. Suppose, for example, that it became 
possible to generalize about how much the rate of learning in operations 
of one type would differ from that in other types of operations; or to ap- 
ply the learning curve to a variety of business operations that could not 
be classified as "manufacturing," but where learning was no less a factor 
in output. 

The challenge before management today is to make this final stage a 
reality. From the beginning of World War II until last year there was, 
except for occasional brief periods, a lack of competition as businessmen 
used to know it. Not unnaturally, executives became less cost-conscious. 
There was no great urgency about spending the time and money that 
would be necessary to develop practicable learning curves such as those 
used in the aircraft industry. But now, with hard competition back, there 
is every incentive to collect the data, plot the curve, test it, refine it. The 
aggressive, forward-looking management will not wait long. The learning 
curve is a device that promises to have many important uses. 




T. V. Houser* 

This article is based on a Charles Coolidge Farlin Memorial Lecture 
give?i under the auspices of the Philadelphia Chapter of the American 
Marketing Association on May 27, 1958. Mr. Houser states that the mar- 
keting executive's knowledge of the market and methods must play a 
major part in areas of the business which traditionally have not been con- 
sidered a part of his responsibility, and that this will result in increased 
business effectiveness. 


Two patterns of distribution have emerged in the United States. The 
traditional pattern is that of the large manufacturer whose products are 
distributed by great numbers of small dealers. There may be independent 
or producer-controlled wholesaler activities included in this system. How- 
ever, it is typified by a combination of large producer and small distribu- 

* From the Journal of Marketing, XXIII, 4 (April, 1959), 363-69. Reprinted by per- 
mission of the author and the Journal of Marketing, national quarterly publication 
of the American Marketing Association. 


Role of the Marketing Executive 363 

tors, with the manufacturer exercising marketing leadership through a 
variety of methods, chiefly advertising, to presell the product and secure 

In competition with this system of large producer and small distributor 
is that of the large distributor furnishing an outlet for the products of a 
great number of small manufacturers. In this case, product determination 
is the responsibility of the distributor; and marketing leadership also comes 
from the distributor through product advertising, display, direct-sales 
effort to the consumer, and customer-credit arrangements. 

Neither of these patterns is static. For example, major producers of 
durable goods have been reaching forward to the distributor function 
... in an effort to integrate the selling and production effort, and to get 
the benefits of increased volume and lower costs. 


The chief executive officer of a business must keep a sense of balance 
between the various functions comprising his organization. He must also 
determine which functional viewpoint will have the greatest influence on 
over-all decisions. Some executive officers give great weight to the ac- 
counting-and-control function. In other companies where the emphasis is 
placed on technical areas, the engineering executive and the production 
executive are considered the most important. 

The view here is that the marketing junction is the one to carry the 
most weight. 

Selling effort starts with the determination of the product. Therefore, 
the marketing function must have a strong voice in research, development, 
and production decisions. 

So long as the customer is free to buy from competing sellers and pro- 
ducers, the function of the business closest to the customer must carry 
weight in influencing the decisions that determine the character and cost 
of the product. Necessarily, this requires marketing executives who are 
sales minded, thoroughly grounded in business economics, sensitive to 
both the possibilities and limitations of mass-production techniques, and 
profit conscious. 



Most firms carry on some form of applied research independent of cur- 
rent production engineering activity. This may involve prospective im- 
provements in quality, utility, or costs of current products, or may be di- 
rected toward new products. 

364 Marketing Operations and the Administrator 

The chief executive officer must decide on the scope of this effort in 
relation to the business as a whole; but the marketing executive should 
have a major voice in the objectives toward which this activity is di- 
rected. Salesmanship of a product really starts at this point. Very often 
the production cost of a desirable feature added to an existing product is 
a more effective expenditure of money than the more conventional ave- 
nues of sales expense. Yet a proper balance between various elements of 
unit costs must be maintained. The marketing executive should help to 
set cost limitations on product-research objectives, so that impractical 
wqrk will not bd undertaken. 

In like manner, research work on new products must face the marketing 
test. All too often the engineering and design department and the produc- 
tion department have ideas as to new products which will fit into the type 
of equipment, skills, and techniques of the work force, but which cannot 
be distributed through the existing marketing channels. Full consideration 
must be given to what is involved in creating appropriate channels of 
distribution for such new products. The marketing executive is the one to 
spell this out. 

Product research usually involves improvement in quality or utility; 
this may mean greater aesthetic satisfactions in use, greater conveniences 
in use, or longer life. Here again the marketing executive should establish 
the priority and relative values, because cost limitations ordinarily will not 
permit the adoption of all desirable improvements. Here he must reflect 
a correct interpretation of the desires of that part of the general market he 
is serving, in order to strike the balance of features and price that will 
expand the market. 


Consider also the competitive conditions which a business faces. A mar- 
keting executive must determine the number of different items which will 
constitute his firm's line of products, be they electric irons, men's shoes, or 
television sets. As a rule, his sales force wants to have a product to match 
every individual offering of all competing manufacturers. On the other 
hand, the engineers and production men know that the smaller the range 
of items, the lower the unit costs will be because of less tooling expense 
and longer production runs. The marketing executive must stand between 
these two pressures, sensitive to each, and make appropriate judgments. 

Any merchandise line usually consists of a beginning number with mini- 
mum quality standards, minimum features, and primary emphasis on price 
appeal ... a medium number designed to appeal to the largest market, 
with recognizable, justifiable, and demonstrable differences of quality and 
features . . . and a "top-of-the-line" number with qualities appealing to 
the higher-income market. 

Role of the Marketing Executive ^65 

The marketing executive must know when to add a number, when to 
drop one, and when to substitute a new one for one to be dropped. The 
"image" which the public forms of a given line of products is largely in 
his hands, not because he originates the ideas expressed in merchandise, but 
because he has a grasp of the economics involved and a feel for the public's 
attitudes and preferences. 


The next step is the specific product, determination of design, and the 
organization of the sequence of production operations. At this stage it be- 
comes necessary for estimates of specific quantities to be made for each 
item in the proposed line. Only the marketing function can have the ex- 
perience and general feel of the market necessary to develop dependable 
estimates. The degree of tooling must be determined, involving the balanc- 
ing of the additional amortization which complete tooling requires against 
the added labor costs where less tooling is possible. 

While the engineering designs and the production plans are being made, 
aesthetic design is also under way. Here again is an area where judgment 
and balance from a marketing viewpoint are essential. All too often cus- 
tomer convenience and utility are sacrificed to a standard of pure aesthetic 
design which exists in the artist's mind. 

Choices of various materials are often possible, as well as various proc- 
essing methods; and the marketing viewpoint must carry weight wherever 
decisions on these points might involve consumer preferences. More than 
this, problems of economical shipment, breakage, mechanical failure, or 
rapid deterioration must be taken into account. 


About this time budgets begin to take shape, based on sales prospects, 
manufacturing schedules, and projections of various elements of expense 
and investment. This is the area of direct sales management, which in- 
cludes the management and control of the sales organization, advertising, 
display, and all sales activities. 


In this connection, two points should be made. 

The first is the importance of co-ordination of advertising, sales effort, 
service of supply, and all the other devices used in modern selling. If these 
parts are not properly fitted together, a great deal of the research time 
and production effort can be dissipated by a faulty sales campaign. 

366 Marketing Operations and the Administrator 

The second point deals with the management of the sales organization. 
This calls for the very best kind of personnel administration. It involves 
not only selection, training, and incentives, but also relationship to man- 
agement. Sound personnel policies are important in every part of any busi- 
ness, but are vital to the sales organization. 


The marketing executive should realize that the intrinsic quality of an 
article of merchandise is determined by its design, the kind of materials 
used in its manufacture, and the degree of precision and competence in its 
manufacture. When it reaches the end of the assembly line, packed for 
shipment, its level of quality is forever established. It may still go through 
a number of shipments and trans-shipments, be unloaded, and held in 
storage a number of times; title of ownership may be transferred several 
times; and it will be handled in different places by different people before 
it reaches its destination in a customer's home. As a rule a relatively small 
part of its ultimate cost to the consumer was expended for material, labor, 
factory overhead, engineering, and tooling costs. For a common run of 
items, this proportion would range around 40 to 50 per cent. 

But the expenses incurred beyond the assembly line, although not add- 
ing to the intrinsic quality of the article, have added greatly to its cost. 
The technology of mass production is so well developed that one can 
hardly expect significant contributions to improved living standards, or 
the creation of broader markets through cost reductions. 

From a mathematical standpoint, value could be defined as equal to 
quality divided by price. Double the quality without change of price, and 
you double the value; double the price without change of quality, and 
you have one-half the former value. Since quality is determined on the 
manufacturer's assembly line and has not changed at the retail level, the 
value of the product to the consumer is reduced by successive distribution 
costs. Since it is at the retail level that the manufacturer's product must 
gain preference of retail customers, any shortcuts or efficiencies intro- 
duced along the line will increase the value of the product in comparison 
with competition. 

Realizing that layer upon layer of distributive costs does not add to the 
worth of his product, the marketing executive turns his attention to ways 
and means of reducing this spread between direct production costs of 
goods and the price the consumer pays. Two functions of distribution are 
so basic that they cannot be avoided. 

The first is the wholesaling or distributor function, which may be car- 
ried out by the manufacturer either at the factory or a number of regional 
points of distribution; or by independent operators of regional distribution 
points; or by a large-scale retailer or operator of multiple stores. 

Role of the Marketing Executive 367 

The other function is the point of contact with the ultimate consumer, 
usually through retail stores (with mail-order a unique exception). Ef- 
forts to bypass these elementary functions through direct-to-consumer 
selling have been limited and successful only with larger margins than per- 
mitted by conventional channels. 


One of the great savings in the cost of producing goods, particularly of 
a seasonal character, is to maintain a stable level of production day in and 
day out. In two plants with comparable individual costs, one plant may 
have markedly higher total costs at the end of a year, even if both have 
adequate volume. Analysis will generally show, in the plant with the lower 
costs, that every machine runs all or most of the year, with individual runs 
of a given part sufficient to justify the cost of machine adjustments to 
make that part. Year-end cost savings of 10 to 15 per cent are obtainable, 
when a highly seasonal article is produced steadily throughout the year. 
This means proper marketing planning. 

The full benefits of this type of production can best be realized when 
there is assurance that the retail outlets will carry the line produced. Un- 
certainty as to stocking the line tends to be removed if firm contracts can 
be established with large retail chains. 


Sometimes marketing executives overlook the importance of having 
correct and adequate information "tags" on articles. The differences be- 
tween various qualities must be made clear to the prospective buyer; this 
is also a means of educating the salesperson on the job. Many times the 
selling effort by the manufacturer has gone on the outside of a container, 
which is thrown away when the article is placed in retail display! 

An example of the power of packaging involves rosebushes. Here is an 
item, pruned root and branch, which may not suggest to the customer its 
ultimate beauty. However, with proper packaging to illustrate its ap- 
pearance in bloom, a rosebush becomes an item with great appeal, and 
produces the plus sales that come from impulse buying by the customer. 


Store display and arrangement are important, and very often proper 
help from the manufacturer is important. However, such help must be 
compatible with the general appearance and character of the store, practi- 
cal in taking care of customers on a busy day. 

Probably the greatest single opportunity for improvement in the whole 
field of distribution lies in stockkeeping. The marketing executive must 

3 68 Marketing Operations and the Administrator 

see the whole sequence from the original design to the product in the 
home, and maintain a steady, consistent flow of goods through the entire 
stream. He must realize that a salesman who sells his product to a retailer 
by inducing that retailer to carry a wider range of selection than his busi- 
ness will support is really doing the manufacturer a great disservice. 

So far as possible, each retailer should be counseled as to just what 
range of products would be most suitable, taking into account sales vol- 
ume and floor-space, so that the store is always in stock and not worried 
about over-stocks. 


A perennial problem is the evaluation of advertising. The objective to 
be accomplished by a given advertisement should be kept clearly in mind. 
For instance, some programs are aimed at preselling the consumer; others 
are psychological weapons in the competitive war, and are intended more 
to convince a retailer that the public is already presold; while still others 
have the more intangible justification of building prestige. 

Some manufacturers of cereal, soap, and similar products set up a unit 
advertising amount in the cost projection of a product and then keep an 
advertising variance account, just as is done in the shop for material, labor, 
or overhead standard costs and variances. Wherever this technique is ap- 
plicable, it sharpens the marketing executive's analysis and appraisal of the 
advertising dollar. 


The decentralization of industry is another matter which requires the 
attention of the marketing executive. 

Decentralization can be the key to developing a balanced economy 
regionally, maintaining more even population growth, and contributing 
to a more stable economy and better standard of living. It is an almost 
ideal situation when an area develops a reasonable balance between agri- 
culture and processing industries. If an area can add raw material for those 
industries, then it becomes that much stronger, with even greater basic 
importance to the economy as a whole. 

So far as marketing is concerned, there is a vast difference between 
selling the products of a distant producer in a given region as against those 
of a local manufacturer. Even if a distributing warehouse is used, the local 
plant still has great advantages. There is something to be said for being an 
integral part of a community. There is the good will which comes from 
participation in local and regional public service . . . the value of close 

Role of the Marketing Executive 369 

contact between the sales personnel and plant executives . . . greater at- 
tention to service of supply and to mechanical service matters . . . 
greater concentration regionally of selling techniques, including adver- 

This does not imply that the marketing executive must be the dominant 
voice in decisions concerning a distant branch plant. But, if a company 
faces the problem of expansion, he should analyze the opportunities for 
increased volume resulting from a regional plant and develop the basic 
economics of such an operation. Usually raw materials are available in part 
in most regions; but, if they are not, the saving in shipping finished goods 
from the heart of a region greatly offsets the costs of shipping raw ma- 
terials to the plant. Unless the marketing executive has the imagination 
and strength of conviction to present fully the branch proposal, the pro- 
duction executive will probably see to it that merely a new wing is added 
to the present plant. 


Next, consider the growing importance of consumer credit. This de- 
velopment has brought a train of problems in its wake; but our great mass- 
production industries could not be supported without credit selling. 

Installment-credit selling in some form is here to stay, although no one 
knows exactly what level it will attain in our national economy. Increas- 
ingly, credit is an important point to take into account in selecting dealers 
and distributors. Where sound financing is not available to otherwise ac- 
ceptable dealers, the manufacturer can probably organize such financing 
services or help arrange for them. 


Ours is a dynamic economy, constantly undergoing rapid changes, and 
these changes can have immediate or long-range effects on marketing 
plans. For example, what weight should be given to the fact that, for 
every person added to the potential labor force since 1940, two were 
added to the group to be supported by the labor force? 

What should be the influence on marketing of the growing numbers 
of our population in the lower-age groups? 

What is the marketing impact of the growth of the suburbs? 

What are the implications for advertising and sales of the movement of 
our population to states in the West and the Southwest, and the population 
growth of other states adjacent to the population centers of the Atlantic 

oy Marketing Operations and the Administrator 

What is the significance to the marketing function of an increase -since 
1950 of 35 per cent in the highly skilled professional and technical group 
in industry, and a decrease in the number of unskilled workers? 

Should an increase of 32 per cent since 1946 in the number of women 
in the work force be a factor in making marketing plans and projections? 

What is the significance of the tremendous gain in productivity on the 
farm, with one farm worker now producing the food requirements of 
twenty people in 1957, as against less than ten in 1940? 

What bearing does increased leisure time have on marketing decisions? 

Public and private insurance plans covering hospital, surgical, and medi- 
cal care, health-and-accident insurance, life insurance, and pension funds 
have assets totaling an estimated %'jiVi billion, and are increasing at the 
rate of almost $5 billion per year. What is the importance of this deferred 
income to long-range sales planning? 

Forces such as these exert a steady pressure for an infinite variety of 
goods. Knowledge of these forces, their directions, and their intensity— 
with an understanding of their effect on business— give the marketing ex- 
ecutive his greatest value. 



Charles W. Smith* 

Marketing as a separate management junction, not to be identified solely 
"with selling and advertising, is regarded by the author as a necessary and 
important step toward improving profits in the future. A marketing de- 
partment so conceived "will be integrated with those of other company 


As a phenomenon of industrial organization, the marketing department 
is a relatively recent development. Fundamentally, it is an outgrowth of 
more than twenty years of discussion about the need to increase distribu- 
tion efficiency. Of all the subjects that might have been selected, I can 
think of no other that would be more timely or important to any group 
of thinking executives. Planned marketing, based on recognition of the 
"total marketing" concept, is gaining increasing recognition in top man- 

* From Planned Marketing—Management 's Responsibility, Marketing for Executive 
Series, No. 4 (1957). Reprinted by permission of Planned Marketing— Management's 
Responsibility and the American Marketing Association. 

Planned Marketing 37 1 

agement circles as one of the principal approaches to improving a com- 
pany's profits. 

Reduced to the simplest possible terms, the marketing department of 
any industrial company is simply an organization device for giving formal 
recognition to the concept that marketing is a separate, and very im- 
portant, management function. Despite its simplicity, this concept has 
proven in many cases to be a difficult one to build into an existing cor- 
porate structure. For whenever any new concept develops in the field 
of scientific management, some companies seem to mistake the form for 
the substance of what is basically a sound idea. 

A well-planned marketing operation cannot be achieved simply by cre- 
ating a marketing department on paper. Major companies that have 
endeavored to establish a really effective marketing department have 
learned that the task involves much more than merely assigning some ex- 
ecutive a new title and drawing up a new organization chart. 

For a marketing department to become an effective part of an enter- 
prise, its activities must be integrated with those of every other company 
department. Those responsible for administering its operations must 
clearly visualize the role of the marketing function as separate and distinct 
from either selling or advertising, while continuing to recognize the vital 
importance of both these functions in the total marketing operation. Top 
company executives must also adjust their patterns of decision making to 
reflect the concept that all sound corporate planning is based on a pre- 
determination of the markets to be served. 

In many companies where the "total marketing" philosophy has not as 
yet gained acceptance, forward plans are still being founded almost en- 
tirely on engineering, production, or financial considerations. As long as 
conditions make it possible for such companies to operate profitably, there 
is no pressure on them to move in the direction of establishing marketing 
departments. Three basic factors in our economy, however, are tending to 
make it increasingly difficult for any company to achieve optimum profit 
results without giving primary consideration to marketing factors in de- 
veloping its long-range plans. 


The first of these factors is the rapidity of change in the economy. Mar- 
kets that once were stable, and continuing, can no longer be taken for 
granted; change has come to be the distinguishing characteristic of our age. 

In such a world of constant and dramatic change, it is becoming more 
and more profitable to conduct the research required to find out every- 
thing that can be predetermined about a potential market before attempt- 
ing to serve it. We can no longer take it for granted that we will always 

372 Marketing Operations and the Ad?ninistrator 

be able to sell a good product if we can make it. Total capacity to produce 
is now so great that demand for a product must be created continuously 
to insure profitable employment of productive facilities. 

First, we must be sure that products are designed to fit the needs of the 
market as well or better than competing lines. We must also make sure we 
have created an awareness of those needs, of sufficient intensity, to bring 
potential consumers into the market to buy. 

When marketing research tells us that there is no ready demand for a 
product, whether it is new or established, we must then determine what 
sales appeals can be used to create demand and locate potential customers 
who will respond to such appeals. Only then do we have a sound basis 
for a long-range promotion program that will support a planned level 
of output. 

Companies that have not taken these basic marketing planning steps, or 
have miscalculated their market planning, have often found themselves 
saddled with beautiful plants that could not be operated profitably. Fi- 
nancial retrenchment then becomes the order of the day, and as competi- 
tive conditions become more severe, companies that have failed to do 
effective marketing planning are among the first to fall. 

Another aspect of our changing world is the constant shifting of con- 
sumer preferences from one product to another. This is most marked in 
the so-called "style industries," which have long recognized the need for 
keeping their fingers constantly on the pulse of market demand. This fac- 
tor of change, however, is rapidly becoming more important in other 
traditionally more stable industries. 

These market changes do not occur overnight. They can be foreseen by 
the alert observer, sensitive to the importance of new developments. Yet 
many companies have gone blindly ahead making products in declining 
demand until their profit positions have been completely eroded. Other 
companies, more alert to market developments, have seized opportunities 
to shift investment into profitable new fields, and rapidly changing market 
conditions have given strong impetus to the development of marketing 


The second major factor in the growing importance of the marketing 
function is the increasing complexity of distribution. For the past twenty- 
five years the primary attention of top management, in a large majority 
of companies, has been concentrated on the critical problems involved in 
raising production efficiency; distribution efficiency has been very largely 
taken for granted. 

One amazing aspect of our industrial economy is the level of enthusiasm 
exhibited by top-management executives for fundamental research in the 
field of engineering. Company after company has poured hundreds of 

Planned Marketing 373 

thousands of dollars into the construction of new scientific research 
laboratories that are now producing a constant stream of new product 
ideas. Relatively few companies, however, have as yet conducted either 
the kinds or the total amount of distribution research required to evaluate 
the product ideas developed by their engineering research. Better co-ordi- 
nation of marketing research and product research efforts is thus becom- 
ing a major problem of top management. Thousands of dollars have been 
wasted in developing products that no one has been able to figure out how 
to market profitably. At the same time, many opportunities for profitable 
application of new product ideas have been overlooked for lack of sound 
distribution research. 

In some industries, such as the home appliance field, distribution has 
become the primary factor determining a company's competitive position. 
The trend towards production of a complete line of appliances by major 
manufacturers has made it increasingly difficult for producers of partial 
lines to secure top-flight wholesale outlets and has contributed to the re- 
cent wave of mergers between partial line manufacturers seeking to main- 
tain their competitive positions in the field. 

More and more products and product lines are being designed and de- 
veloped in terms of the competitive conditions imposed by the channels 
of distribution through which they must flow. Pricing and promotion 
policies are also being adjusted to reflect trade practices of distribution 
outlets over which the manufacturer has only limited control. 

Distribution systems that once consisted of only one basic channel 
have now been expanded to cover several channels. In the surgical dress- 
ings industry, for instance, the traditional pattern of distribution through 
drug wholesalers has been amplified by direct sales to variety store and 
supermarket chain warehouses. Changes in the distribution systems used 
by manufacturers have come about as a result of three developments: (a) 
efforts of established retail outlets to broaden the base of their operations 
by the addition of new product lines, i.e., the creation of drug sundry de- 
partments in food stores; (b) growth of new types of outlets, i.e., super- 
markets and discount houses; and (c) efforts of manufacturers to broaden 
the base of their operations by the addition of new product lines, i.e., the 
entry of plumbing-heating supply manufacturers into the kitchen cabinet 
and appliance field. 

As a result of these changes, problems involved in planning and con- 
trolling inventories held in the distribution pipeline have also become in- 
creasingly complex. Product lines are longer and customer service require- 
ments are more demanding. Warehousing space and labor costs and 
working capital tied up in obsolete or slow-moving inventories are all 
rising rapidly where distribution costs have not been under tight control. 
Analysis of situations in which distribution costs were found to be out of 
line has disclosed inefficiencies that are often the result of a poorly de- 
signed distribution system. 

374 Marketing Operations and the Administrator 

Any company which subjects its distribution operations to searching 
analysis will usually find many opportunities for profit improvement such 
as: warehouses poorly located from the standpoint of fast, economical 
service; archaic order-handling systems; greater use of modern materials 
handling equipment and improved methods of inventory control; profita- 
ble orders lost to competition through inability to give prompt reliable 
service, or lack of sufficient sales contacts. 

The increasing complexity of distribution systems has placed a premium 
on sound planning and control of distribution operations. Typical func- 
tions that can be assigned to marketing departments include: setting stand- 
ards of performance for every important distribution function; evaluating 
methods and procedures to determine ways of increasing efficiency; and 
designing and locating facilities with a view to long-term growth potential. 


The third major factor that has given impetus to the creation of mar- 
keting departments is the increasing importance of sales forecasts as the 
basis for all corporate profit planning operations. Capital investment re- 
quirements of corporate enterprise are now so great that top-management 
executives must have sound marketing plans on which to base judgments 
regarding the potential profitability of future operations. This is particu- 
larly true of industries in which automation has become an important 
factor determining a company's cost and profit position. 

Many companies are now projecting their operations as far as ten to 
twenty years into the future. No one can forecast in any detail what will 
take place so far in the future, but decisions made today regarding the 
scale and location of plants and warehouses actually set effective limits 
on a company's competitive position and profit structure for many years. 
Technological developments are coming so fast that only the most astute 
management executives can keep them in proper perspective. The more 
information a company has about long-term growth patterns the better 
its capital investment plans will be. 


The vital role of marketing planning can perhaps best be appreciated 
in terms of the profit planning cycle of a make-and-sell enterprise shown 
in Chart i. The cycle starts with the long- and short-term sales fore- 
casts and the sales plan. It is the responsibility of marketing departments to 
prepare the sales forecasts and sales-plan figures which in turn provide 
the basis for all subsequent steps in the profit planning cycle. 

The kind of marketing planning described is formal in nature and the 
need for such formal planning seems to depend on the size and complexity 
of the individual enterprise. 

Planned Marketing 




O 1 

z q 

z § 

Z -D 

5 B 

.— — ^. 

< u 
co z 

■u Q 

= 03 


>; co z 



Q ,_ 

O uj 



o «° 


t 5 








° X ^ 
X «" * 


O ™ i_ 

oo t 

Q Z Q 

^ tu -3 

x > m 




Z Q 







3 S '' x f N ! 


> i- ; 



== oe s- 



5 < < 


Z =i iu 

• O > Q 

u 2 <a 


O — Q 



U Z 



ET— ^ J 

Lu U 
^ UJ 

3 Q t 

^ § o 

_j < Q 

a- ^ 3 

< * 




P U 



376 Marketing Operations and the Administrator 

All companies— even the smallest— carry on informal planning activities 
which are simply taken for granted. The need for more formal planning 
methods is usually not recognized until operations reach a certain stage 
of complexity. At that point, informal plans no longer provide a sound 
basis for administration of the enterprise. 

The concept of integrated formal planning of a company's operations 
grew out of the work of the managerial accountants who developed the 
basic concept that the manager of any phase of a company's operations 
should be held accountable for the results he has achieved against pre- 
determined objective standards of accomplishment. Budgeting and fore- 
casting grew out of this concept. The cycle of planning illustrated in the 
diagram indicates the ultimate development of this accounting idea as 
applied under the method of direct costing. Thus it is apparent that the 
marketing planning functions is a by-product, or an outgrowth, of the 
long-term trend toward the use of more scientific management methods. 

The growth of the marketing planning function has been slow in some 
companies because top executives have failed to recognize that it requires 
a different type of skill than that required for success in selling or adver- 
tising. The experience of most companies has been that formal marketing 
plans were not well prepared until this responsibility was clearly recog- 
nized as a separate and distinct management function and assigned to a 
marketing department. 

Planned marketing is a key that management can use to open the door 
to increased profits; for this reason more and more companies will be 
establishing marketing departments in the years immediately ahead. 


Five top executives tell how cost accounting is or can be used in market- 
ing—its limitations, advantages, and prospects. 


By E. J. Hanley 

Allegheny Ludlum 
Steel Corp. 

* From Industrial Marketing, XLII, 12 (1957), 102, 104, 106. Copyright 1957 by 
Advertising Publications, Inc., 200 East Illinois Street, Chicago 11, Illinois. Reprinted 
by permission of the publisher. 

Cost Accounting in Marketing 377 

Cost accounting, properly interpreted, is essential to management for 
guidance in avoiding unprofitable sales and seeking to make sales which 
produce optimum profits. But cost accounting itself cannot be the whole 

Management must also take into account such marketing and production 
considerations as the relationship of the unprofitable product to the more 
profitable ones, the advantages of carrying a "full line," the stage of mar- 
ket development of a particular product or line, the effect of volume on 
costs, the maintenance of a stable work force and the most effective em- 
ployment of production equipment. 

Not every sale can be expected to result in maximum profits. A reliable 
business must stand ready to meet its customers' needs, and in so doing it 
must be prepared to make deliveries of some products that are less profita- 
ble than others. 

But the more we know about our over-all cost picture and the better 
data we have on the real costs of producing and selling each product in 
each marketing area, the better we are able to aim our market develop- 
ment in the direction of better profitability. 

No cost figures are really "exact," and this is inclined to be particularly 
true of marketing costs. Nevertheless, it should be possible for most com- 
panies to allocate on a reasonable basis their over-all marketing costs to 
individual products, territories and even customers. The extent to which 
it is practical and desirable to do so may depend, however, upon the cost 
of obtaining the necessary data and the value placed upon the breakdowns 
when they are obtained. 

New data processing systems and equipment are being brought into 
operation to enable management to do a better job of collecting and an- 
alyzing the information needed for correct and more precise allocation of 
costs. Such machines and systems will not supplant sound business judg- 
ment, but by proper coding and tabulation, make the information more 
useful in arriving at accurate cost breakdowns, correct judgments and 
sound marketing policies. 

In these days of rising costs and keen competition, management cannot 
afford not to take advantage of every modern tool and technique that it 
can bring to bear on the problems of accurate cost accounting and sound 
market analysis. 


By C. S. Hallauer 
Bausch & Lomb 
Optical Co. 
Rochester, N. Y. 

378 Marketing Operations and the Administrator 

Competent and comprehensive budgeting in every phase of our opera- 
tions has always seemed to us to be an indispensable part of modern man- 
agement and operating control. 

Our controller's division works closely with our different sales depart- 
ments in setting up our annual budget. It also helps in periodical re- 
appraisals. It provides a constant flow of information on sales vs. budget. 
It shows us— in monthly reports— which lines are affording a satisfactory 
profit and which lines are unsatisfactory. 

With this information, the sales and advertising departments can empha- 
size those products which offer the best profit potential. Their ability to 
do so has a direct relationship to our operational profit, even though it is 
not always practical to discontinue or to completely de-emphasize un- 
profitable lines. 

The information provided by our controller's division is comprehensive 
in character and adequately covers the primary needs of management and 
our sales divisions. We know our marketing costs on every one of our 
many sales classifications. We have a complete geographical breakdown 
for selling expense of our regional sales divisions. We know specifically the 
cost of operations in every one of more than 1 50 branches and the degree 
to which each contributes to our over-all accomplishment in terms of 
both sales and profits. 

This sort of information serves to spotlight areas of strength and of 
weakness. We have found it most valuable in the formulation of future 


By R. A. Pritzker 
Great American 
Elyria, O. 

Cost analysis of the sales function for a company is essential to good 
management. This axiom is, however, not the whole story. The real prob- 
lem, which requires management thinking, is the economic feasibility of 
obtaining such costs. 

Several years ago while engaged in consulting work, I had occasion to 
show a midwestern manufacturer that for a number of years his marketing 
cost in the New York area actually exceeded his gross sales. Although this 
is a striking and obvious example, the executive involved has been tre- 
mendously successful in sales. In a medium or large corporation where 

Cost Accounting in Marketing 379 

the sale of many products is involved, it is almost common to find un- 
profitable sales efforts that could be eliminated with proper cost analysis. 

With modern electronic tabulating equipment, it is possible to produce 
almost any conceivable breakdown of sales expense. Most companies must, 
however, carefully determine what information is necessary and can be 
the catalyst for corrective action. It is difficult to generalize on the amount 
of breakdown that is feasible because of the variation in product from 
company to company. For example, a firm whose average invoice is less 
than $2 would find marketing statistics or marketing cost analysis far more 
expensive than would a business whose average invoice is $ i ,000. 

The Colson Corp. (a subsidiary of Great American Industries) has 
found invaluable its analysis of marketing expenses. In certain areas the 
company uses distributors, while in others, direct salesmen are employed. 
This gives rise to the question of which method is more economical and 
which produces the most sales. Although there are many opinions ex- 
pressed on this question, we have found by a careful analysis that is is a 
complex question and must be handled differently in each marketing area. 

It is apparent, then, that generalization is impossible regarding the use- 
fulness of cost breakdowns. It is necessary to first make an economic study 
comparing the expense of a cost breakdown to the probable increased 
profits resulting from such a program. Most companies find, I think, that 
certain cost analyses give great returns, while other breakdowns are far 
more expensive than the profits received. 


By R. B. Seymour 
Loven Chemical 
of California 
Neivhall, Cal. 

Like other management tools, cost accounting must be used for the 
control of all phases of marketing programs. The principal pitfalls are 
usually in research and development which are sometimes considered "off 
limits." Yet, cost accounting must be applied at least as rigorously here as 
in advertising and sales promotion. Otherwise, results will be unrealistic 
and of questionable value. 

In every well-managed sales organization, marketing costs can be deter- 
mined on the basis of product line, territories and customers. The precise- 
ness of the data depends on factors which are not always under control. 
Nevertheless, a realistic breakdown usually provides surprisingly good and 
sometimes embarrassing answers. 

^8 Marketing Operations and the Administrator 

Certainly the data can be sufficiently quantitative to permit manage- 
ment to be objective and make decisions essential for progress. It must be 
emphasized, however, that judgment and intelligent interpretation are at 
least equal in importance to the exactness of these data. 


By F. H. Orbison 


Appleton Woolen Mills 

Appleton, Wisconsin 

Is cost accounting the answer to avoiding unprofitable sales, if used for 
all aspects of a company's marketing program— research, product develop- 
ment, advertising, sales management and sales promotion? 

There is no such thing as "the answer." 

The question you have asked is: "How can cost accounting best be used 
in managing sales?" There appear to be two possibilities. 

i. For the control of both direct and indirect selling costs. This is very 
difficult since arbitrary judgments are involved in establishing the basis 
for allocations. 

2. For measuring the effectiveness of allied programs such as research, 
product development, advertising, sales promotion and sales management. 
This seems quite worth while. 

The determination of values is a responsibility of good management. 
In this determination, cost accounting can be a helpful tool, but figures 
don't have feelings. The evaluation of people is more important than the 
evaluation of figures. People produce the figures and they are measured 
by them. 

In avoiding unprofitable sales, we don't look to cost accounting as being 
the answer. It is a means to reaching an answer but not the end. 

Do you believe it is possible for the average company to break down 
its marketing costs, exactly, by any or all of the following classifications: 
individual products, territories, customers? 

We do not believe it is possible for any company to break down its 
marketing costs "exactly" by individual products, territories or by cus- 

This is because cost accounting of some sales costs can only be based 
on arbitrary allocations. For example: 

i. How can you break down travel expenses by products or by cus- 
tomers? This is done best by territories, assuming you have a single 
method of distribution for all your products. 

2. How can the cost of samples be broken down by customers? We 
can do this by products— perhaps even by territories. 

Cost Accounting in Marketing 381 

3. How can entertainment costs be broken down by products? We can 
do it by customers and perhaps by territories but not by all three. 

There are probably other variations for other companies. It is important 
to remember that each company is different from the average. 

For us, a breakdown of costs by territories is the best basis for alloca- 
tion. But even so, there are some costs which are not readily allocable by 
territory— research, product development, advertising, and to some extent 
sales promotion and sales management. 

To try to allocate these costs by territory would produce unreliable 
or misleading figures. Marketing decisions based on figures alone would 
be ill-advised. 





Thomas J. Wright* 

The author tells how a personnel quality control program can be set up, 
what it should cover, how to make it function, what departments within 
a company should handle various parts, and briefly explains how such an 
over-all program can be effectively operated. 

The control of quality is a most important factor in the manufacture of 
material products; so is quality control important with respect to the 
human beings who are engaged in the various activities involved in the 
manufacturing process and in the related processes of design, inspection, 
sales, and so forth. The effectiveness of the techniques developed for 
controlling the quality of the products can be minimized or increased to 
a considerable extent depending upon the program which has been de- 
veloped for personnel quality control. 

The objective of this article, is to indicate how a program for personnel 
quality control functions, and how the techniques developed for the 

* From Advanced Management, XVIII, 10 (1953), 17-20. Reprinted by permission 

of Advanced Management. 

3 8z 

Fro gram for Personnel Quality Control 383 

control of product quality may be applied to the personnel field. The steps 
to be covered involve the following items: job specification, screening 
techniques, induction to company, training on the job, job performance 
review, and organization to accomplish the foregoing. 

The examples and forms included in this paper are drawn for the most 
part from a large and progressive carpet manufacturing company. This 
company, employing some 10,000 production and maintenance, technical, 
sales, clerical and sales personnel has made considerable progress during 
recent years in the development and application of personnel quality 

The job specification serves the same purpose with respect to the 
control of the quality of personnel that the material specification serves 
insofar as material quality control is concerned. That is, it specifies the 
duties to be performed under specified working conditions by the in- 
cumbent of a specific job along with the related responsibilities and lists 
the characteristics such as previous formal education, specialized skills, 
on-the-job training, job test scores, and so forth, which the incumbent 
must measure up to in order to perform the job satisfactorily. 

The personnel requisition serves the same purpose with respect to the 
recruiting of the proper personnel for a job that the purchase order serves 
when material supplies are involved. Often a copy of the job specification 
is attached or, if not, a digest of the information contained thereon is 
used. In addition, the number of persons required, the date required, 
location of work, and starting rate are included plus provision for the 
necessary approvals which must be obtained before recruiting starts. 
Copies of the personnel requisition usually are sent to the payroll, per- 
sonnel and requisitioning departments after the selection has been made. 

Just as all material shipped into a well-managed plant is inspected before 
being accepted, so must new personnel be subjected to a similar procedure. 
This personnel acceptance inspection usually consists of the following 
parts: interview, tests, medical examination and references. 


The job applicant usually is requested to fill out an application for 
employment, thereby furnishing information regarding formal education, 
special training, work experience, age, marital status, etc., on the basis of 
which the initial screening may be made to determine whether he meets 
the job specification requirements sufficiently to warrant an interview. 
If the job is of a technical or artistic nature, the applicant is often required 
to furnish samples of his work when filing his application. This is similar 
to the manufacturer who is required to furnish samples of his product 
when he is trying to obtain an order either for the first time or for a large 
quantity from a purchaser. 

3 8 4 

Personnel Control and the Administrator 

If a review of the application by the interviewer indicates that the 
applicant apparently possesses the necessary job qualifications, he is given 
an interview, during which any required personnel tests are given. During 
the interview an attempt is made to analyze the applicant's personal traits, 
appearance, ambitions, etc., to determine whether he will "fit" into the 
company's working organization, assuming that he has the other necessary 

In addition to the pre-employment tests the applicant is usually required 
to serve a probationary period of from one week to several months, de- 
pending on the nature of the job, during which time he works under close 
observation to determine whether he can perform the required duties. 

In order to check on the accuracy of the job applicant's statements re- 
garding his previous education and experience, reference letters are sent 
by the employer to the applicant's previous employers and heads of 
schools which he attended. The replies to these letters, particularly from 
the previous employers, often contain little other than recorded informa- 
tion regarding length of employment and job classification. Of course, 
if the applicant's record was a very good one and his former employer is 
desirous of helping him make a change, the reply may contain favorable 

An important step in the control of the quality of personnel which most 
of the larger employers are utilizing is the pre-employment and periodic 
medical examination. The former is usually given at the conclusion of 
the employment interview and personnel testing procedure. 


While there are a number of diseases which immediately disqualify a 
person for employment, this is not true with respect to physical handicaps 
such as the loss of a leg, an eye, defective hearing, etc. There are many 
jobs on which a handicapped person can often perform better than one 
who has all of his faculties. It is the duty of the medical department to fit 
the person to the job from a physical standpoint while at the same time 
using discretion in providing for future replacements in the better jobs 
by selecting a high percentage of fully-qualified (physically) employees. 

The practice of giving periodic physical examinations to detect any 
incipient diseases is becoming more popular, particularly in companies 
which are large enough to support a full-time medical staff with modern 
diagnostic equipment. 

It has been found from experience that the quality of a company's 
personnel is improved if, before the new employee is actually assigned to 
a specific job, he is given an opportunity of orienting himself by learning 

Fro gram for Personnel Quality Control 385 

something about the company's general personnel and other policies, its 
products, organization and manufacturing and distributing facilities. This 
indoctrination is usually a responsibility of the personnel department's 
training division and is accomplished by the distribution of pamphlets 
and handbooks, talks by selected management personnel and visits to 
company plants and showrooms whenever feasible. 

The induction program, in addition to giving the new employee a 
bird's-eye view of the company of which he is becoming a part also pro- 
vides an opportunity of providing him with first-hand information and 
answers to questions instead of waiting until he is faced with a situation 
requiring an answer which may be given him by someone who is not 
qualified to give him the correct information. 


Years ago, before industrial training programs reached their present 
stature, the employee was assigned to a job, given a few brief instructions 
by the foreman and then was left to shift for himself. When he encoun- 
tered a problem which made it impossible for him to proceed, he asked 
his foreman for assistance, or, if the foreman was not available— as was 
often the case— turned to a fellow-worker for aid. While in either case 
there was a considerable loss of time involved, the latter practice was 
particularly bad since the new employee picked up many of the poor 
work habits which the older employee had developed over the years. 

Today most progressive companies provide for a training period ranging 
from a day to several weeks in length, depending upon the job difficulty, 
during which the new employee receives instruction in the proper method 
of doing the job from either an experienced operator- trainer or a full-time 
trainer. The former is to be preferred when the number of new employees 
fluctuates because the trainer when not training is a regular operator, 
thus keeping the overhead cost down. This on-the-job training is particu- 
larly important when the employee is working under an incentive plan or 
operating an expensive piece of machinery where standard production 
must be reached within a short period in order to bring the employee's 
earnings up to par and keep the overhead cost per unit within proper 

From a quality-control standpoint, the learning of correct work habits 
and the proper steps to take in performing the job, as well as how to 
avoid the making of scrap, is a most important factor in so far as personnel 
is concerned. This training must be repeated in many cases when a 
periodical review indicates that the employee has deviated from the pro- 
cedures which he was taught to follow originally. 

386 Personnel Control and the Administrator 


The principles of quality control are followed with respect to personnel 
in the technique of job performance review more so than in any of the 
other techniques covered by this paper with the possible exception of 
employee testing. Just as samples of the manufactured products are in- 
spected by means of gauges or other measuring instruments, at intervals 
to determine whether the process is holding its own with respect to the 
specified quality limits, so is the performance of the individuals who are 
engaged in the various operations evaluated at regular intervals for similar 
reasons. This review of personnel job performance is commonly referred 
to as merit rating, merit review, performance review, and so forth. While 
in many respects, as with other phases of quality control, this technique 
is still in its infancy, it is found today in most of the progressive companies. 

Job performance review, as this technique will be referred to in this 
article, serves the following three major purposes: (1) To provide an 
indication of merited salary increases; (2) To indicate in what areas 
additional training is required; and (3) To indicate which employees are 
promotional prospects. 

While the principle applies to jobs in all levels of the organization, in 
practice the incumbents of "white collar" jobs have their performance 
reviewed in more companies than do those of manual or "blue collar" jobs. 
This is probably due to a considerable extent to the fact that manual 
workers are usually paid on a piece rate basis or on a single day-rate basis 
as compared with the salary ranges (from minimum to maximum) which 
exist for "white collar jobs." In addition, the fact that today a large 
proportion of the manual workers in large manufacturing and service 
companies are unionized has tended to limit the application of the job 
performance review technique insofar as this type of employee is con- 
cerned, with the result that seniority has replaced ability as the principal 
criterion of promotability in many companies. 


An employee's job performance is usually reviewed by his immediate 
supervisor, who is the person best qualified to do this. This review is 
usually checked by the person to whom the supervisor reports in order 
to obtain at least two separate points of view. This practice may be com- 
pared to the checking of the inspectors' work by the check inspector in 
so far as manufactured products are concerned. 

Just as the inspectors in the plant are furnished with gauges and measur- 
ing instruments to check the quality of the manufactured product, the 

Program for Personnel Quality Control 3 8 7 

job performance reviewers, or raters as they are often referred to, are 
furnished with rating forms to assist them in determining to what extent 
the employees under their direction are fulfilling the requirements of their 
jobs. It should be borne in mind that such forms should be tailor-made for 
the particular company, or even subdivision of a company, in which they 
are to be used. Otherwise, the situation of trying to fit the round peg 
into the square hole arises and the technique of merit is blamed for the 
poor results instead of the unskilled technicians. 

The training of the raters is a must if valid judgments of employees' 
performance are to be obtained. This training must be done at the time the 
rater is first introduced to the technique and periodically thereafter, pref- 
erably immediately preceding or following the time at which the ratings 
are made. Regardless of how good the rating form is, it is the raters who 
in the last analysis make or break the performance review program. 

Insofar as the timing of ratings is concerned, there is no hard and fast 
policy which applies under all circumstances. For example, it is often 
found desirable to rate all of the members of a large clerical group at the 
same time in order to take advantage of the opportunity for comparing 
the ratings of personnel performing the same jobs, that is, typists, billers, 
telephone operators, thus afforded. On the other hand, it is usually found 
more desirable to rate the performance of executive personnel on the an- 
niversary date of their employment in their present positions in view of 
the fact that it is very difficult to compare their jobs with each other. 

With respect to the scoring of personnel performance reviews, the 
same differences of opinion exist among the "experts" with respect to 
numerical versus non-numerical or profile scoring as exists in the related 
technique of job evaluation. Suffice to say that, as Dr. Juan points out 
under Systems of Rating Quality of Product in his book Management of 
Inspection and Quality Control, committees will often spend long hours 
debating such points whereas this item is only a minor one of the many 
decisions which must be made in any quality rating plan. It has been the 
writer's experience in working with both types of systems in personnel 
rating that when a numerical scoring system is used, there is a tendency 
for the raters and ratees to concentrate on the over-all numerical "score" 
and to neglect the individual factors score which indicates the ratee's 
strength and weakness-this is particularly true where the ratee's salary 
depends upon the result of the performance review. 

Just as it is important that when inspectors find a defective product 
being manufactured they call it to the attention of the operators, so 
should the rater discuss the ratings of his employees with them as soon 
as they have been returned to him by the reviewer. This discussion of 
the ratings between the rater and ratee is one of the most important, if 

Personnel Control and the Administrator 

not the most important, phases of the performance review program and is 
probably the one which is often handled most ineffectually, principally, 
perhaps, because of the difficulty of checking it. As a matter of fact, many 
companies which have performance review plans in effect make no pro- 
vision for the discussion of the ratings with the individual rated. 


After the rating forms have been returned to a central source, usually 
the personnel department, when the reviews with the employees rated 
have been completed, the results of the ratings are analyzed, for several 
reasons. In the first place, a chart of over-all ratings is prepared to show 
the relationship between the ratings of employees on similar or related 
jobs. The degree of agreement between the actual distribution of the 
rating results and the well-known "bell" curve of normal distribution is a 
measure of the validity of the rating scales or forms. In most cases in 
actual practice, perfect agreement is not found because of distortions of 
the samples due to employment screening, training, smallness of groups, 

Similar charts of each rater's results are often prepared where the 
number of employees involved is sufficiently great in order to determine 
whether there is a reasonable amount of distinction among the employees' 
performance indicated. A definite slew to the right usually indicates a 
high rater while one to the left indicates the opposite, in either of which 
case further training of the rater is indicated. 


In addition to charts of the over-all ratings, similar charts are usually 
prepared for the ratings on each separate performance factor, such as 
accuracy, cooperation, job knowledge, etc., in order to determine 
whether they are furnishing significant results. Such an analysis of factors 
provided significant data when a performance review form for supervisory 
jobs was developed. In this case, an analysis of the rating results (numerical 
scores from weighted factors) obtained in two ratings of several hundred 
supervisors showed that a 6o-factor form could be reduced to eight factors 
and eight questions without reducing the effectiveness of the rating, by 
eliminating insignificant factors and combining others which differed to a 
very slight extent. 

Whereas a centralized group of trained quality control technicians are 

Fro gram for Personnel Quality Control 389 

needed to develop, co-ordinate, and analyze results of material quality con- 
trol programs, a similar group of personnel technicians perform the same 
function in the field of personnel administration. In the latter case, how- 
ever, the cooperation of the line organization is probably even more 
important in order to assure the success of the program than it is in the 

While practically every person in the personnel department in a posi- 
tion of any importance has some responsibility for personnel quality con- 
trol, the actual development and administration of the techniques described 
in the foregoing sections of this paper are usually the responsibility of the 
following groups: employment, medical, training, and wage and salary. 

Obviously, the screening techniques are carried out by the employment 
group along with induction whereas the medical group handles medical 
examinations and the training group handles training. Because of the fact 
that the results of a job performance review are usually used as a criterion 
in connection with the granting of wage increases, the responsibility for 
the administration of these reviews is often included in the wage and 
salary group. However, because of the close relationship between this 
technique and those of testing, training, placement, etc., the groups 
handling these functions review the results of the reviews and collaborate 
in any changes that are made. 

The foregoing sections of this article dealt primarily with a comparison 
between quality control with respect to manufactured products and with 
respect to personnel and with the techniques used for achieving the latter. 
It is evident, therefore, that quality control principles are applicable in 
many fields. 

One important factor which should be kept uppermost in mind by 
those engaged in any phase of personnel quality control or by those con- 
templating the initiating of any of the techniques involved is that they 
are dealing with human beings rather than with machines or other in- 
animate objects. Human nature being what it is, there are no two persons 
exactly alike and their differences must be recognized much more so than 
minor differences in material things. Just -as quality-mindedness must be 
instilled in the minds of all employees of a company, regardless of posi- 
tion, so must human-relations-mindedness be similarly instilled. 

Much of the responsibility for developing this attitude insofar as per- 
sonnel quality control is concerned, rests with the company's personnel 
department since it is in a position to exert the greatest single influence on 
the company's personnel in this respect. If a company whose top manage- 
ment is really human-relations minded develops a competent staff of per- 
sonnel technicians and gives them the necessary authority commensurable 
with their responsibility, results in the form of improved quality of 
personnel should be evident in a relatively short time. 

390 Personnel Control and the Ad?ninistrator 


Chris Argyris* 

Budgets are utilized as pressure devices for increasing efficiency but 
at the same time tend to generate forces which in the long run decrease 
efficiency. The various conflicts, misunderstandings, and cross-purposes 
that develop between persons involved in various phases of the budgetary 
process are discussed at some length based on a survey in depth of three 

Budgets are accounting techniques designed to control costs through 
people. As such their impact is felt by everyone in the organization. They 
are continuously being brought into the picture when anyone is trying 
to determine, plan, and implement an organizational policy or practice. 
Moreover, budgets frequently serve as a basis for rewarding and penaliz- 
ing those in the organization. Failure to meet the budget in many plants 
invites much punishment; success, much reward. 

Because budgets affect people so directly, it seems appropriate to ask 
ourselves some questions about them. What are the effects of budgets on 
the human relationships in the organization? How well are budgets ac- 
complishing their practical purposes? How can the use of budgets be 
improved to make them more effective? This article reports some of the 
results of a pilot study designed to suggest answers to questions like these. 


To keep the research problem within manageable limits, it was decided 
to focus primarily on the effects of manufacturing budgets (e.g., produc- 
tion, waste, and error budgets) upon the front-line supervisor. To this end 
our research group made a field study of three small plants (i.e., with less 
than 1,500 employees) manufacturing both "custom made" products and 
products "for stock"— all three of them unionized and covering the full 
range from highly skilled to nonskilled workers. (None of the plants, it 

* From Harvard Business Review, XXXI, 1 (1953), 97-110. Reprinted by permis- 
sion of the Harvard Business Review. 

Author's Note: The research which led to this article was carried out under the 
auspices of the Controllership Foundation. Dr. Schuyler D. Hoslett, Director of Ex- 
ecutive Program in Business Administration, Columbia University, was in over-all 
charge of the project. Mr. Frank B. Miller of Cornell University assisted me. The 
entire research is reported in The Impact of Budgets on People (New York, Con- 
trollership Foundation, 1952). 

Human Problems with Budgets 391 

should be noted, has a supervisory incentive system as part of its budget 
system.) In one plant we interviewed a 100% sample of front-line super- 
visors, in another a 90% sample, and in the third a 25% sample, using 
non-directive questions; and in addition we observed many of the same 
supervisors in action. 

Just in case my observations suggest that I have little sympathy for 
budgets, let me say that any such impression stems from the negative 
reactions which the interviews themselves produced. This is perhaps in- 
evitable, when one keeps the following points about budgets in mind: 

1. Budgets are, first of all, evaluation instruments. Because they tend to set 
goals against which to measure people, they naturally are complained about. 

2. Budgets are one of the few evaluation processes that are always in writ- 
ing and therefore concrete. Thus, some of the supervisors tend to use budgets 
as "whipping posts" in order to release their feelings about many other (often 
totally unrelated) problems. 

3. Budgets are thought of as pressure devices. As such they produce the 
same kind of unfavorable reactions as do other kinds of pressure regardless of 
origin. In fact, the analysis I shall make of the effects of management pressure 
upon supervisors is not necessarily limited to budgets. For example, a company 
"saddled" with a domineering executive but which has no budget may well be 
affected by the same factors as those reported herein. 


The process of preparing a manufacturing budget is much the same 
in all plants: 

It usually starts with a meeting of the controller, the assistant controller, and 
a group of top-management members to determine over-all financial goals for 
the company in the forthcoming year. The controller's staff then translates the 
financial goals into the detailed breakdowns required for departmental budgets. 
This preliminary budget is then sent to all superintendents, who are asked to 
scrutinize it carefully and report any alterations they wish to make. 

During their period of scrutiny, the superintendents (middle management) 
discuss the budget with their own supervisory group in a series of meetings. 
The first-line foremen do not usually take part in these discussions, although 
their ideas are requested when the superintendents deem it desirable. 

Once the superintendents have their budget modifications clearly in mind, 
a meeting is held with the controller and his staff. Both parties come to the 
meeting "armed to the teeth" with "ammunition" to back their demands. After 
the disagreements are resolved, all parties sign the new budget proposal. The 
superintendents return to their offices, awaiting the new budget and the ex- 
pected drive to "put it over." 

So much for background information. Now, let us turn to a discussion 
of some of the effects budgets seem to have upon people. This discussion 
will be limited to five major areas: (1) the problem resulting from the 
fact that budgets are used as a pressure device, (2) the problem of the 
budget supervisor's success and the factory supervisor's failure, (3) the 

^g 2 Personnel Control and the Administrator 

problem of department-centered supervisors, (4) the problem resulting 
from the fact that budgets are used as a medium for personality expres- 
sion, and (5) a case example of human problems related to budgets. In 
conclusion, some lines of action for management to follow in meeting 
these problems will be suggested. 


One of the most common of the factory supervisors' assumptions about 
budgets is that they can be used as a pressure device to increase produc- 
tion efficiency. Finance people also admit to the attitude that budgets 
help "keep employees on the ball" by raising their goals and increasing 
their motivation. The problem of the effects of pressure applied through 
budgets seems to be at the core of the budget problem. 


Employees believe that pressure from above is due to top management's 
assumption that most workers are basically or inherently lazy. Employees 
also believe that top management thinks the workers do not have enough 
motivation of their own to do the best possible job. And first-line super- 
visors have the same beliefs. 

Interviews with top-management officials revealed that these employee 
beliefs were not totally unfounded, as a few quotations from some of the 
top management (both line and finance) make clear: 

"I'll tell you my honest opinion. About 5% of the people work, 10% of the 
people think they work, and the other 85% would rather die than work!" 

"I think there is a definite need for more pressure. People have to be 
needled a bit. Man is inherently lazy, and if we could only increase the pres- 
sure, I think the budget system would be more effective." 

"There are lots of workers in this plant, hundreds of them, who don't have 
the capacity to do things other than what they're doing. And they're lazy! 
They might be able to develop some capacities, although I think there are a 
lot of them who couldn't even if they wanted to. But they don't even have 
the desire." 

Such feelings, even if never openly expressed to the employees, filter 
through to them in very subtle ways. Once they sense that feelings of 
this kind exist in top management, they may become very resentful. Some 
supervisors, who apparently felt that budgets reflected this situation, ex- 
pressed warnings like the following: 

"The employees have an established conception of a fair output. Now, if 
you believe, like most financial people do, that the average worker is out to 
cheat the company all the time, I give up." 

"Managements ought to change their attitudes that employees are out to get 

"Well, I guess they think that we need this needling. But / don't think so." 

Human Problems with Budgets 393 


How do people react to pressure? In the three plants studied factory 
supervisors felt they were working under pressure and that the budget 
was the principal instrument of pressure. Management exerts pressure on 
the work force in many ways, of course; budgets are but one way. Being 
concrete, however, budgets seem to serve as a medium through which the 
total effects of management pressure are best expressed. As such they 
become an excellent focal point for studying the effect of pressure on 
people in a working organization. 

To help clarify what happens when management "puts on the pressure" 
let us think of a situation in specific terms: 

For the sake of illustration, assume that the technical efficiency of a plant is at 
a maximum and that human efficiency, or effort, has to be increased if produc- 
tion is to be raised. The work force of the plant is producing at 70% of its 
efficiency. Since the level of efficiency is 70%, there must be certain forces 
which keep it from falling below 70% and certain forces which prevent it 
from going above 70%. 

Some of the forces which keep it from falling below 70% are: (a) budget 
talks to foremen; (b) red circling of poor showing of the departments in the 
budget results; (c) production drives; (d) pep talks by supervisors using 
budgets as evidence; (e) threat of reprimand if the budget is not met; and 
(f ) threat of feelings of failure if budget is not met. 

Some of the forces which prevent efficiency from going above 70% are: 
(a) informal agreements among employees not to produce more; (b) fear of 
loss of job if efficiency increases; (c) foreman (union) agreements against 
"speedups"; (d) abilities of individual employees; (e) abilities of work teams 
as a who